In [14]:
# using openCV to read the image and plt to show the image
plt.imshow(cv2.imread(brain_df.mask_path[623]))
Out[14]:
<matplotlib.image.AxesImage at 0x169bd1a2ed0>
# Task: Improve the speed and accuracy of detecting and localising Brain tumor based on MRI scans
# Which will drastically remove the cost of cancer diagnosis and help in early diagnoses of brain tumors
# develop a model to detect and localise Brain tumor
# You have been provided with 3900 brain MRI scans along with their brain tumor location
# Processes: Using AI to improve diseases detection and location process
# Build and train a segmentation Res-U-Net model to localise brain tumor in images
# Talk about AI in health care
# Deep learning
# for the training data, I have two images; the brain MRI (to show if it has cancer or not) and a mask associated with the image.
# First stage we are going to train a ResNet deeplearning classifier (Residual Neural Network) model
# Second stage, if a tumor has been detected after classification, then we need to localize the tumor using ResUNet Segmentation Model; it tells the location of the tumor
# Function of the image segmentation is to understand and extract information from images at the pixel-level
pip install opencv-python
Requirement already satisfied: opencv-python in c:\users\olayi\anaconda3\lib\site-packages (4.9.0.80) Requirement already satisfied: numpy>=1.21.2 in c:\users\olayi\anaconda3\lib\site-packages (from opencv-python) (1.24.3) Note: you may need to restart the kernel to use updated packages.
# Importing important libraries
import pandas as pd
import numpy as np
import seaborn as sns
import matplotlib.pyplot as plt
import zipfile
import cv2
import tensorflow as tf
from skimage import io
from tensorflow.python.keras import Sequential
from tensorflow.keras import layers, optimizers
from tensorflow.keras.applications import DenseNet121
from tensorflow.keras.applications.resnet50 import ResNet50
from tensorflow.keras.layers import *
from tensorflow.keras.models import Model, load_model
from tensorflow.keras.initializers import glorot_uniform
from tensorflow.keras.utils import plot_model
from tensorflow.keras.callbacks import ReduceLROnPlateau, EarlyStopping, ModelCheckpoint, LearningRateScheduler
from IPython.display import display
from tensorflow.keras import backend as k
from sklearn.preprocessing import StandardScaler, normalize
import os
import glob
import random
%matplotlib inline
from warnings import filterwarnings
filterwarnings('ignore')
WARNING:tensorflow:From C:\Users\olayi\anaconda3\Lib\site-packages\keras\src\losses.py:2976: The name tf.losses.sparse_softmax_cross_entropy is deprecated. Please use tf.compat.v1.losses.sparse_softmax_cross_entropy instead.
C:\Users\olayi\anaconda3\Lib\site-packages\paramiko\transport.py:219: CryptographyDeprecationWarning: Blowfish has been deprecated "class": algorithms.Blowfish,
# Define the path to the downloads directory
downloads_path = os.path.join(os.path.expanduser('~'), 'Downloads')
# Check if the directory exists
if os.path.exists(r"C:\Users\olayi\Downloads\Projects\Health AI\Brain_Tumor\Healthcare+AI+Datasets\Healthcare AI Datasets\Brain_MRI"):
print("Downloads directory found at:", downloads_path)
else:
print("Downloads directory not found!")
# Change directory to the downloads folder
os.chdir(r"C:\Users\olayi\Downloads\Projects\Health AI\Brain_Tumor\Healthcare+AI+Datasets\Healthcare AI Datasets\Brain_MRI")
Downloads directory found at: C:\Users\olayi\Downloads
# Loading the dataset that contain Brain MRI and their corresponding mask
# loaded the mask path
# mask can be represented by associating pixel values with their coordinates
brain_df = pd.read_csv(r"C:\Users\olayi\Downloads\Projects\Health AI\Brain_Tumor\Healthcare+AI+Datasets\Healthcare AI Datasets\Brain_MRI\data_mask.csv")
brain_df.info()
<class 'pandas.core.frame.DataFrame'> RangeIndex: 3929 entries, 0 to 3928 Data columns (total 4 columns): # Column Non-Null Count Dtype --- ------ -------------- ----- 0 patient_id 3929 non-null object 1 image_path 3929 non-null object 2 mask_path 3929 non-null object 3 mask 3929 non-null int64 dtypes: int64(1), object(3) memory usage: 122.9+ KB
brain_df.head()
| patient_id | image_path | mask_path | mask | |
|---|---|---|---|---|
| 0 | TCGA_CS_5395_19981004 | TCGA_CS_5395_19981004/TCGA_CS_5395_19981004_1.tif | TCGA_CS_5395_19981004/TCGA_CS_5395_19981004_1_... | 0 |
| 1 | TCGA_CS_5395_19981004 | TCGA_CS_4944_20010208/TCGA_CS_4944_20010208_1.tif | TCGA_CS_4944_20010208/TCGA_CS_4944_20010208_1_... | 0 |
| 2 | TCGA_CS_5395_19981004 | TCGA_CS_4941_19960909/TCGA_CS_4941_19960909_1.tif | TCGA_CS_4941_19960909/TCGA_CS_4941_19960909_1_... | 0 |
| 3 | TCGA_CS_5395_19981004 | TCGA_CS_4943_20000902/TCGA_CS_4943_20000902_1.tif | TCGA_CS_4943_20000902/TCGA_CS_4943_20000902_1_... | 0 |
| 4 | TCGA_CS_5395_19981004 | TCGA_CS_5396_20010302/TCGA_CS_5396_20010302_1.tif | TCGA_CS_5396_20010302/TCGA_CS_5396_20010302_1_... | 0 |
# Percentage of healthy sampples( without tumor)
# which means we have unbalanced dataset
len(brain_df[brain_df['mask'] == 0])/len(brain_df)
0.6505472130313057
# counts of healthy samples and unhealthy samples
brain_df['mask'].value_counts()
0 2556 1 1373 Name: mask, dtype: int64
# To know the number of categories we have in the mask
brain_df['mask'].value_counts().index
Int64Index([0, 1], dtype='int64')
# using plotly to show the categories
# plotly shows interactive bar chart
import plotly.graph_objects as go
fig = go.Figure([go.Bar(x = brain_df['mask'].value_counts().index, y = brain_df['mask'].value_counts())])
fig.update_traces(marker_color = 'rgb(0,200,0)', marker_line_color = 'rgb(0,255,0)', marker_line_width = 3, opacity = 0.5)
fig.update_layout(title='Distribution of Categories',
xaxis_title='Categories',
yaxis_title='Count')
fig.show()
# mask path
brain_df['mask_path']
0 TCGA_CS_5395_19981004/TCGA_CS_5395_19981004_1_...
1 TCGA_CS_4944_20010208/TCGA_CS_4944_20010208_1_...
2 TCGA_CS_4941_19960909/TCGA_CS_4941_19960909_1_...
3 TCGA_CS_4943_20000902/TCGA_CS_4943_20000902_1_...
4 TCGA_CS_5396_20010302/TCGA_CS_5396_20010302_1_...
...
3924 TCGA_HT_A61B_19991127/TCGA_HT_A61B_19991127_86...
3925 TCGA_HT_A61A_20000127/TCGA_HT_A61A_20000127_87...
3926 TCGA_HT_A61B_19991127/TCGA_HT_A61B_19991127_87...
3927 TCGA_HT_A61A_20000127/TCGA_HT_A61A_20000127_88...
3928 TCGA_HT_A61B_19991127/TCGA_HT_A61B_19991127_88...
Name: mask_path, Length: 3929, dtype: object
# Image path
brain_df['image_path']
0 TCGA_CS_5395_19981004/TCGA_CS_5395_19981004_1.tif
1 TCGA_CS_4944_20010208/TCGA_CS_4944_20010208_1.tif
2 TCGA_CS_4941_19960909/TCGA_CS_4941_19960909_1.tif
3 TCGA_CS_4943_20000902/TCGA_CS_4943_20000902_1.tif
4 TCGA_CS_5396_20010302/TCGA_CS_5396_20010302_1.tif
...
3924 TCGA_HT_A61B_19991127/TCGA_HT_A61B_19991127_86...
3925 TCGA_HT_A61A_20000127/TCGA_HT_A61A_20000127_87...
3926 TCGA_HT_A61B_19991127/TCGA_HT_A61B_19991127_87...
3927 TCGA_HT_A61A_20000127/TCGA_HT_A61A_20000127_88...
3928 TCGA_HT_A61B_19991127/TCGA_HT_A61B_19991127_88...
Name: image_path, Length: 3929, dtype: object
# using openCV to read the image and plt to show the image
plt.imshow(cv2.imread(brain_df.mask_path[623]))
<matplotlib.image.AxesImage at 0x169bd1a2ed0>
plt.imshow(cv2.imread(brain_df.image_path[623]))
<matplotlib.image.AxesImage at 0x169bd1d3310>
# visualizing the images(MRI and mask) in the dataset
fig, axis = plt.subplots(6, 2, figsize = (16,25))
count = 0
for i in range(6):
i = random.randint(0, len(brain_df)) # Select a random index
axis[count][0].title.set_text('Brain MRI') # Set title
axis[count][0].imshow(cv2.imread(brain_df.image_path[i])) # MRI
axis[count][1].title.set_text('Mask - ' + str(brain_df['mask'][i])) # Plot title on the mask (0 or 1)
axis[count][1].imshow(cv2.imread(brain_df.mask_path[i]))
count += 1
fig.tight_layout()
# visualizing the 12 randomly selected MRI and it correspong Mask from only unhealthy patients
# showing the MRI and the ,mask on top of each other (showing the mask in red color)
fig, axis = plt.subplots(12, 3, figsize = (16,40))
count = 0
for i in range(len(brain_df)):
if brain_df['mask'][i] == 1 and count <12:
img = io.imread(brain_df.image_path[i])
axis[count][0].title.set_text('Brain MRI')
axis[count][0].imshow(img)
mask = io.imread(brain_df.mask_path[i])
axis[count][1].title.set_text('mask')
axis[count][1].imshow(mask, cmap = 'gray')
img[mask == 255] = (255,0,0)
axis[count][2].title.set_text('MRI with Mask')
axis[count][2].imshow(img)
count += 1
fig.tight_layout()
# Drop columns not needed for the training
brain_df_train = brain_df.drop(['patient_id'], axis =1)
brain_df_train.shape
(3929, 3)
# Convert the mask column to a categorical format.
brain_df_train['mask'] = brain_df_train['mask'].apply(lambda x : str(x))
brain_df_train.info()
<class 'pandas.core.frame.DataFrame'> RangeIndex: 3929 entries, 0 to 3928 Data columns (total 3 columns): # Column Non-Null Count Dtype --- ------ -------------- ----- 0 image_path 3929 non-null object 1 mask_path 3929 non-null object 2 mask 3929 non-null object dtypes: object(3) memory usage: 92.2+ KB
# Split the data into train and test data
from sklearn.model_selection import train_test_split
train, test = train_test_split(brain_df_train, test_size = 0.15)
# Create a image generator to generate images in batch format and do some processing on the data
from keras.preprocessing.image import ImageDataGenerator
# Create a data generator which scales the data from 0 to 1 and makes validation split of 0.15
datagen = ImageDataGenerator(rescale = 1./255., validation_split = 0.15)
train_generator = datagen.flow_from_dataframe(
dataframe = train,
directory = './',
x_col = 'image_path',
y_col = 'mask',
subset='training',
batch_size = 16,
shuffle = True,
class_mode ='categorical',
target_size = (256, 256))
val_generator = datagen.flow_from_dataframe(
dataframe = train,
directory = './',
x_col = 'image_path',
y_col = 'mask',
subset='validation',
batch_size = 16,
shuffle = True,
class_mode ='categorical',
target_size = (256, 256))
# Create a data generator for the test images
test_datagen = ImageDataGenerator(rescale = 1./255.)
test_generator = datagen.flow_from_dataframe(
dataframe = test,
directory = './',
x_col = 'image_path',
y_col = 'mask',
batch_size = 16,
shuffle = False,
class_mode ='categorical',
target_size = (256, 256))
Found 2839 validated image filenames belonging to 2 classes. Found 500 validated image filenames belonging to 2 classes. Found 590 validated image filenames belonging to 2 classes.
# Using an already pre-trained model (Transfer learning)
# creating the ResNet50 base model
basemodel = ResNet50(weights ='imagenet', include_top = False, input_tensor= Input(shape=(256, 256, 3)))
WARNING:tensorflow:From C:\Users\olayi\anaconda3\Lib\site-packages\keras\src\backend.py:1398: The name tf.executing_eagerly_outside_functions is deprecated. Please use tf.compat.v1.executing_eagerly_outside_functions instead. WARNING:tensorflow:From C:\Users\olayi\anaconda3\Lib\site-packages\keras\src\layers\normalization\batch_normalization.py:979: The name tf.nn.fused_batch_norm is deprecated. Please use tf.compat.v1.nn.fused_batch_norm instead.
basemodel.summary()
Model: "resnet50"
__________________________________________________________________________________________________
Layer (type) Output Shape Param # Connected to
==================================================================================================
input_1 (InputLayer) [(None, 256, 256, 3)] 0 []
conv1_pad (ZeroPadding2D) (None, 262, 262, 3) 0 ['input_1[0][0]']
conv1_conv (Conv2D) (None, 128, 128, 64) 9472 ['conv1_pad[0][0]']
conv1_bn (BatchNormalizati (None, 128, 128, 64) 256 ['conv1_conv[0][0]']
on)
conv1_relu (Activation) (None, 128, 128, 64) 0 ['conv1_bn[0][0]']
pool1_pad (ZeroPadding2D) (None, 130, 130, 64) 0 ['conv1_relu[0][0]']
pool1_pool (MaxPooling2D) (None, 64, 64, 64) 0 ['pool1_pad[0][0]']
conv2_block1_1_conv (Conv2 (None, 64, 64, 64) 4160 ['pool1_pool[0][0]']
D)
conv2_block1_1_bn (BatchNo (None, 64, 64, 64) 256 ['conv2_block1_1_conv[0][0]']
rmalization)
conv2_block1_1_relu (Activ (None, 64, 64, 64) 0 ['conv2_block1_1_bn[0][0]']
ation)
conv2_block1_2_conv (Conv2 (None, 64, 64, 64) 36928 ['conv2_block1_1_relu[0][0]']
D)
conv2_block1_2_bn (BatchNo (None, 64, 64, 64) 256 ['conv2_block1_2_conv[0][0]']
rmalization)
conv2_block1_2_relu (Activ (None, 64, 64, 64) 0 ['conv2_block1_2_bn[0][0]']
ation)
conv2_block1_0_conv (Conv2 (None, 64, 64, 256) 16640 ['pool1_pool[0][0]']
D)
conv2_block1_3_conv (Conv2 (None, 64, 64, 256) 16640 ['conv2_block1_2_relu[0][0]']
D)
conv2_block1_0_bn (BatchNo (None, 64, 64, 256) 1024 ['conv2_block1_0_conv[0][0]']
rmalization)
conv2_block1_3_bn (BatchNo (None, 64, 64, 256) 1024 ['conv2_block1_3_conv[0][0]']
rmalization)
conv2_block1_add (Add) (None, 64, 64, 256) 0 ['conv2_block1_0_bn[0][0]',
'conv2_block1_3_bn[0][0]']
conv2_block1_out (Activati (None, 64, 64, 256) 0 ['conv2_block1_add[0][0]']
on)
conv2_block2_1_conv (Conv2 (None, 64, 64, 64) 16448 ['conv2_block1_out[0][0]']
D)
conv2_block2_1_bn (BatchNo (None, 64, 64, 64) 256 ['conv2_block2_1_conv[0][0]']
rmalization)
conv2_block2_1_relu (Activ (None, 64, 64, 64) 0 ['conv2_block2_1_bn[0][0]']
ation)
conv2_block2_2_conv (Conv2 (None, 64, 64, 64) 36928 ['conv2_block2_1_relu[0][0]']
D)
conv2_block2_2_bn (BatchNo (None, 64, 64, 64) 256 ['conv2_block2_2_conv[0][0]']
rmalization)
conv2_block2_2_relu (Activ (None, 64, 64, 64) 0 ['conv2_block2_2_bn[0][0]']
ation)
conv2_block2_3_conv (Conv2 (None, 64, 64, 256) 16640 ['conv2_block2_2_relu[0][0]']
D)
conv2_block2_3_bn (BatchNo (None, 64, 64, 256) 1024 ['conv2_block2_3_conv[0][0]']
rmalization)
conv2_block2_add (Add) (None, 64, 64, 256) 0 ['conv2_block1_out[0][0]',
'conv2_block2_3_bn[0][0]']
conv2_block2_out (Activati (None, 64, 64, 256) 0 ['conv2_block2_add[0][0]']
on)
conv2_block3_1_conv (Conv2 (None, 64, 64, 64) 16448 ['conv2_block2_out[0][0]']
D)
conv2_block3_1_bn (BatchNo (None, 64, 64, 64) 256 ['conv2_block3_1_conv[0][0]']
rmalization)
conv2_block3_1_relu (Activ (None, 64, 64, 64) 0 ['conv2_block3_1_bn[0][0]']
ation)
conv2_block3_2_conv (Conv2 (None, 64, 64, 64) 36928 ['conv2_block3_1_relu[0][0]']
D)
conv2_block3_2_bn (BatchNo (None, 64, 64, 64) 256 ['conv2_block3_2_conv[0][0]']
rmalization)
conv2_block3_2_relu (Activ (None, 64, 64, 64) 0 ['conv2_block3_2_bn[0][0]']
ation)
conv2_block3_3_conv (Conv2 (None, 64, 64, 256) 16640 ['conv2_block3_2_relu[0][0]']
D)
conv2_block3_3_bn (BatchNo (None, 64, 64, 256) 1024 ['conv2_block3_3_conv[0][0]']
rmalization)
conv2_block3_add (Add) (None, 64, 64, 256) 0 ['conv2_block2_out[0][0]',
'conv2_block3_3_bn[0][0]']
conv2_block3_out (Activati (None, 64, 64, 256) 0 ['conv2_block3_add[0][0]']
on)
conv3_block1_1_conv (Conv2 (None, 32, 32, 128) 32896 ['conv2_block3_out[0][0]']
D)
conv3_block1_1_bn (BatchNo (None, 32, 32, 128) 512 ['conv3_block1_1_conv[0][0]']
rmalization)
conv3_block1_1_relu (Activ (None, 32, 32, 128) 0 ['conv3_block1_1_bn[0][0]']
ation)
conv3_block1_2_conv (Conv2 (None, 32, 32, 128) 147584 ['conv3_block1_1_relu[0][0]']
D)
conv3_block1_2_bn (BatchNo (None, 32, 32, 128) 512 ['conv3_block1_2_conv[0][0]']
rmalization)
conv3_block1_2_relu (Activ (None, 32, 32, 128) 0 ['conv3_block1_2_bn[0][0]']
ation)
conv3_block1_0_conv (Conv2 (None, 32, 32, 512) 131584 ['conv2_block3_out[0][0]']
D)
conv3_block1_3_conv (Conv2 (None, 32, 32, 512) 66048 ['conv3_block1_2_relu[0][0]']
D)
conv3_block1_0_bn (BatchNo (None, 32, 32, 512) 2048 ['conv3_block1_0_conv[0][0]']
rmalization)
conv3_block1_3_bn (BatchNo (None, 32, 32, 512) 2048 ['conv3_block1_3_conv[0][0]']
rmalization)
conv3_block1_add (Add) (None, 32, 32, 512) 0 ['conv3_block1_0_bn[0][0]',
'conv3_block1_3_bn[0][0]']
conv3_block1_out (Activati (None, 32, 32, 512) 0 ['conv3_block1_add[0][0]']
on)
conv3_block2_1_conv (Conv2 (None, 32, 32, 128) 65664 ['conv3_block1_out[0][0]']
D)
conv3_block2_1_bn (BatchNo (None, 32, 32, 128) 512 ['conv3_block2_1_conv[0][0]']
rmalization)
conv3_block2_1_relu (Activ (None, 32, 32, 128) 0 ['conv3_block2_1_bn[0][0]']
ation)
conv3_block2_2_conv (Conv2 (None, 32, 32, 128) 147584 ['conv3_block2_1_relu[0][0]']
D)
conv3_block2_2_bn (BatchNo (None, 32, 32, 128) 512 ['conv3_block2_2_conv[0][0]']
rmalization)
conv3_block2_2_relu (Activ (None, 32, 32, 128) 0 ['conv3_block2_2_bn[0][0]']
ation)
conv3_block2_3_conv (Conv2 (None, 32, 32, 512) 66048 ['conv3_block2_2_relu[0][0]']
D)
conv3_block2_3_bn (BatchNo (None, 32, 32, 512) 2048 ['conv3_block2_3_conv[0][0]']
rmalization)
conv3_block2_add (Add) (None, 32, 32, 512) 0 ['conv3_block1_out[0][0]',
'conv3_block2_3_bn[0][0]']
conv3_block2_out (Activati (None, 32, 32, 512) 0 ['conv3_block2_add[0][0]']
on)
conv3_block3_1_conv (Conv2 (None, 32, 32, 128) 65664 ['conv3_block2_out[0][0]']
D)
conv3_block3_1_bn (BatchNo (None, 32, 32, 128) 512 ['conv3_block3_1_conv[0][0]']
rmalization)
conv3_block3_1_relu (Activ (None, 32, 32, 128) 0 ['conv3_block3_1_bn[0][0]']
ation)
conv3_block3_2_conv (Conv2 (None, 32, 32, 128) 147584 ['conv3_block3_1_relu[0][0]']
D)
conv3_block3_2_bn (BatchNo (None, 32, 32, 128) 512 ['conv3_block3_2_conv[0][0]']
rmalization)
conv3_block3_2_relu (Activ (None, 32, 32, 128) 0 ['conv3_block3_2_bn[0][0]']
ation)
conv3_block3_3_conv (Conv2 (None, 32, 32, 512) 66048 ['conv3_block3_2_relu[0][0]']
D)
conv3_block3_3_bn (BatchNo (None, 32, 32, 512) 2048 ['conv3_block3_3_conv[0][0]']
rmalization)
conv3_block3_add (Add) (None, 32, 32, 512) 0 ['conv3_block2_out[0][0]',
'conv3_block3_3_bn[0][0]']
conv3_block3_out (Activati (None, 32, 32, 512) 0 ['conv3_block3_add[0][0]']
on)
conv3_block4_1_conv (Conv2 (None, 32, 32, 128) 65664 ['conv3_block3_out[0][0]']
D)
conv3_block4_1_bn (BatchNo (None, 32, 32, 128) 512 ['conv3_block4_1_conv[0][0]']
rmalization)
conv3_block4_1_relu (Activ (None, 32, 32, 128) 0 ['conv3_block4_1_bn[0][0]']
ation)
conv3_block4_2_conv (Conv2 (None, 32, 32, 128) 147584 ['conv3_block4_1_relu[0][0]']
D)
conv3_block4_2_bn (BatchNo (None, 32, 32, 128) 512 ['conv3_block4_2_conv[0][0]']
rmalization)
conv3_block4_2_relu (Activ (None, 32, 32, 128) 0 ['conv3_block4_2_bn[0][0]']
ation)
conv3_block4_3_conv (Conv2 (None, 32, 32, 512) 66048 ['conv3_block4_2_relu[0][0]']
D)
conv3_block4_3_bn (BatchNo (None, 32, 32, 512) 2048 ['conv3_block4_3_conv[0][0]']
rmalization)
conv3_block4_add (Add) (None, 32, 32, 512) 0 ['conv3_block3_out[0][0]',
'conv3_block4_3_bn[0][0]']
conv3_block4_out (Activati (None, 32, 32, 512) 0 ['conv3_block4_add[0][0]']
on)
conv4_block1_1_conv (Conv2 (None, 16, 16, 256) 131328 ['conv3_block4_out[0][0]']
D)
conv4_block1_1_bn (BatchNo (None, 16, 16, 256) 1024 ['conv4_block1_1_conv[0][0]']
rmalization)
conv4_block1_1_relu (Activ (None, 16, 16, 256) 0 ['conv4_block1_1_bn[0][0]']
ation)
conv4_block1_2_conv (Conv2 (None, 16, 16, 256) 590080 ['conv4_block1_1_relu[0][0]']
D)
conv4_block1_2_bn (BatchNo (None, 16, 16, 256) 1024 ['conv4_block1_2_conv[0][0]']
rmalization)
conv4_block1_2_relu (Activ (None, 16, 16, 256) 0 ['conv4_block1_2_bn[0][0]']
ation)
conv4_block1_0_conv (Conv2 (None, 16, 16, 1024) 525312 ['conv3_block4_out[0][0]']
D)
conv4_block1_3_conv (Conv2 (None, 16, 16, 1024) 263168 ['conv4_block1_2_relu[0][0]']
D)
conv4_block1_0_bn (BatchNo (None, 16, 16, 1024) 4096 ['conv4_block1_0_conv[0][0]']
rmalization)
conv4_block1_3_bn (BatchNo (None, 16, 16, 1024) 4096 ['conv4_block1_3_conv[0][0]']
rmalization)
conv4_block1_add (Add) (None, 16, 16, 1024) 0 ['conv4_block1_0_bn[0][0]',
'conv4_block1_3_bn[0][0]']
conv4_block1_out (Activati (None, 16, 16, 1024) 0 ['conv4_block1_add[0][0]']
on)
conv4_block2_1_conv (Conv2 (None, 16, 16, 256) 262400 ['conv4_block1_out[0][0]']
D)
conv4_block2_1_bn (BatchNo (None, 16, 16, 256) 1024 ['conv4_block2_1_conv[0][0]']
rmalization)
conv4_block2_1_relu (Activ (None, 16, 16, 256) 0 ['conv4_block2_1_bn[0][0]']
ation)
conv4_block2_2_conv (Conv2 (None, 16, 16, 256) 590080 ['conv4_block2_1_relu[0][0]']
D)
conv4_block2_2_bn (BatchNo (None, 16, 16, 256) 1024 ['conv4_block2_2_conv[0][0]']
rmalization)
conv4_block2_2_relu (Activ (None, 16, 16, 256) 0 ['conv4_block2_2_bn[0][0]']
ation)
conv4_block2_3_conv (Conv2 (None, 16, 16, 1024) 263168 ['conv4_block2_2_relu[0][0]']
D)
conv4_block2_3_bn (BatchNo (None, 16, 16, 1024) 4096 ['conv4_block2_3_conv[0][0]']
rmalization)
conv4_block2_add (Add) (None, 16, 16, 1024) 0 ['conv4_block1_out[0][0]',
'conv4_block2_3_bn[0][0]']
conv4_block2_out (Activati (None, 16, 16, 1024) 0 ['conv4_block2_add[0][0]']
on)
conv4_block3_1_conv (Conv2 (None, 16, 16, 256) 262400 ['conv4_block2_out[0][0]']
D)
conv4_block3_1_bn (BatchNo (None, 16, 16, 256) 1024 ['conv4_block3_1_conv[0][0]']
rmalization)
conv4_block3_1_relu (Activ (None, 16, 16, 256) 0 ['conv4_block3_1_bn[0][0]']
ation)
conv4_block3_2_conv (Conv2 (None, 16, 16, 256) 590080 ['conv4_block3_1_relu[0][0]']
D)
conv4_block3_2_bn (BatchNo (None, 16, 16, 256) 1024 ['conv4_block3_2_conv[0][0]']
rmalization)
conv4_block3_2_relu (Activ (None, 16, 16, 256) 0 ['conv4_block3_2_bn[0][0]']
ation)
conv4_block3_3_conv (Conv2 (None, 16, 16, 1024) 263168 ['conv4_block3_2_relu[0][0]']
D)
conv4_block3_3_bn (BatchNo (None, 16, 16, 1024) 4096 ['conv4_block3_3_conv[0][0]']
rmalization)
conv4_block3_add (Add) (None, 16, 16, 1024) 0 ['conv4_block2_out[0][0]',
'conv4_block3_3_bn[0][0]']
conv4_block3_out (Activati (None, 16, 16, 1024) 0 ['conv4_block3_add[0][0]']
on)
conv4_block4_1_conv (Conv2 (None, 16, 16, 256) 262400 ['conv4_block3_out[0][0]']
D)
conv4_block4_1_bn (BatchNo (None, 16, 16, 256) 1024 ['conv4_block4_1_conv[0][0]']
rmalization)
conv4_block4_1_relu (Activ (None, 16, 16, 256) 0 ['conv4_block4_1_bn[0][0]']
ation)
conv4_block4_2_conv (Conv2 (None, 16, 16, 256) 590080 ['conv4_block4_1_relu[0][0]']
D)
conv4_block4_2_bn (BatchNo (None, 16, 16, 256) 1024 ['conv4_block4_2_conv[0][0]']
rmalization)
conv4_block4_2_relu (Activ (None, 16, 16, 256) 0 ['conv4_block4_2_bn[0][0]']
ation)
conv4_block4_3_conv (Conv2 (None, 16, 16, 1024) 263168 ['conv4_block4_2_relu[0][0]']
D)
conv4_block4_3_bn (BatchNo (None, 16, 16, 1024) 4096 ['conv4_block4_3_conv[0][0]']
rmalization)
conv4_block4_add (Add) (None, 16, 16, 1024) 0 ['conv4_block3_out[0][0]',
'conv4_block4_3_bn[0][0]']
conv4_block4_out (Activati (None, 16, 16, 1024) 0 ['conv4_block4_add[0][0]']
on)
conv4_block5_1_conv (Conv2 (None, 16, 16, 256) 262400 ['conv4_block4_out[0][0]']
D)
conv4_block5_1_bn (BatchNo (None, 16, 16, 256) 1024 ['conv4_block5_1_conv[0][0]']
rmalization)
conv4_block5_1_relu (Activ (None, 16, 16, 256) 0 ['conv4_block5_1_bn[0][0]']
ation)
conv4_block5_2_conv (Conv2 (None, 16, 16, 256) 590080 ['conv4_block5_1_relu[0][0]']
D)
conv4_block5_2_bn (BatchNo (None, 16, 16, 256) 1024 ['conv4_block5_2_conv[0][0]']
rmalization)
conv4_block5_2_relu (Activ (None, 16, 16, 256) 0 ['conv4_block5_2_bn[0][0]']
ation)
conv4_block5_3_conv (Conv2 (None, 16, 16, 1024) 263168 ['conv4_block5_2_relu[0][0]']
D)
conv4_block5_3_bn (BatchNo (None, 16, 16, 1024) 4096 ['conv4_block5_3_conv[0][0]']
rmalization)
conv4_block5_add (Add) (None, 16, 16, 1024) 0 ['conv4_block4_out[0][0]',
'conv4_block5_3_bn[0][0]']
conv4_block5_out (Activati (None, 16, 16, 1024) 0 ['conv4_block5_add[0][0]']
on)
conv4_block6_1_conv (Conv2 (None, 16, 16, 256) 262400 ['conv4_block5_out[0][0]']
D)
conv4_block6_1_bn (BatchNo (None, 16, 16, 256) 1024 ['conv4_block6_1_conv[0][0]']
rmalization)
conv4_block6_1_relu (Activ (None, 16, 16, 256) 0 ['conv4_block6_1_bn[0][0]']
ation)
conv4_block6_2_conv (Conv2 (None, 16, 16, 256) 590080 ['conv4_block6_1_relu[0][0]']
D)
conv4_block6_2_bn (BatchNo (None, 16, 16, 256) 1024 ['conv4_block6_2_conv[0][0]']
rmalization)
conv4_block6_2_relu (Activ (None, 16, 16, 256) 0 ['conv4_block6_2_bn[0][0]']
ation)
conv4_block6_3_conv (Conv2 (None, 16, 16, 1024) 263168 ['conv4_block6_2_relu[0][0]']
D)
conv4_block6_3_bn (BatchNo (None, 16, 16, 1024) 4096 ['conv4_block6_3_conv[0][0]']
rmalization)
conv4_block6_add (Add) (None, 16, 16, 1024) 0 ['conv4_block5_out[0][0]',
'conv4_block6_3_bn[0][0]']
conv4_block6_out (Activati (None, 16, 16, 1024) 0 ['conv4_block6_add[0][0]']
on)
conv5_block1_1_conv (Conv2 (None, 8, 8, 512) 524800 ['conv4_block6_out[0][0]']
D)
conv5_block1_1_bn (BatchNo (None, 8, 8, 512) 2048 ['conv5_block1_1_conv[0][0]']
rmalization)
conv5_block1_1_relu (Activ (None, 8, 8, 512) 0 ['conv5_block1_1_bn[0][0]']
ation)
conv5_block1_2_conv (Conv2 (None, 8, 8, 512) 2359808 ['conv5_block1_1_relu[0][0]']
D)
conv5_block1_2_bn (BatchNo (None, 8, 8, 512) 2048 ['conv5_block1_2_conv[0][0]']
rmalization)
conv5_block1_2_relu (Activ (None, 8, 8, 512) 0 ['conv5_block1_2_bn[0][0]']
ation)
conv5_block1_0_conv (Conv2 (None, 8, 8, 2048) 2099200 ['conv4_block6_out[0][0]']
D)
conv5_block1_3_conv (Conv2 (None, 8, 8, 2048) 1050624 ['conv5_block1_2_relu[0][0]']
D)
conv5_block1_0_bn (BatchNo (None, 8, 8, 2048) 8192 ['conv5_block1_0_conv[0][0]']
rmalization)
conv5_block1_3_bn (BatchNo (None, 8, 8, 2048) 8192 ['conv5_block1_3_conv[0][0]']
rmalization)
conv5_block1_add (Add) (None, 8, 8, 2048) 0 ['conv5_block1_0_bn[0][0]',
'conv5_block1_3_bn[0][0]']
conv5_block1_out (Activati (None, 8, 8, 2048) 0 ['conv5_block1_add[0][0]']
on)
conv5_block2_1_conv (Conv2 (None, 8, 8, 512) 1049088 ['conv5_block1_out[0][0]']
D)
conv5_block2_1_bn (BatchNo (None, 8, 8, 512) 2048 ['conv5_block2_1_conv[0][0]']
rmalization)
conv5_block2_1_relu (Activ (None, 8, 8, 512) 0 ['conv5_block2_1_bn[0][0]']
ation)
conv5_block2_2_conv (Conv2 (None, 8, 8, 512) 2359808 ['conv5_block2_1_relu[0][0]']
D)
conv5_block2_2_bn (BatchNo (None, 8, 8, 512) 2048 ['conv5_block2_2_conv[0][0]']
rmalization)
conv5_block2_2_relu (Activ (None, 8, 8, 512) 0 ['conv5_block2_2_bn[0][0]']
ation)
conv5_block2_3_conv (Conv2 (None, 8, 8, 2048) 1050624 ['conv5_block2_2_relu[0][0]']
D)
conv5_block2_3_bn (BatchNo (None, 8, 8, 2048) 8192 ['conv5_block2_3_conv[0][0]']
rmalization)
conv5_block2_add (Add) (None, 8, 8, 2048) 0 ['conv5_block1_out[0][0]',
'conv5_block2_3_bn[0][0]']
conv5_block2_out (Activati (None, 8, 8, 2048) 0 ['conv5_block2_add[0][0]']
on)
conv5_block3_1_conv (Conv2 (None, 8, 8, 512) 1049088 ['conv5_block2_out[0][0]']
D)
conv5_block3_1_bn (BatchNo (None, 8, 8, 512) 2048 ['conv5_block3_1_conv[0][0]']
rmalization)
conv5_block3_1_relu (Activ (None, 8, 8, 512) 0 ['conv5_block3_1_bn[0][0]']
ation)
conv5_block3_2_conv (Conv2 (None, 8, 8, 512) 2359808 ['conv5_block3_1_relu[0][0]']
D)
conv5_block3_2_bn (BatchNo (None, 8, 8, 512) 2048 ['conv5_block3_2_conv[0][0]']
rmalization)
conv5_block3_2_relu (Activ (None, 8, 8, 512) 0 ['conv5_block3_2_bn[0][0]']
ation)
conv5_block3_3_conv (Conv2 (None, 8, 8, 2048) 1050624 ['conv5_block3_2_relu[0][0]']
D)
conv5_block3_3_bn (BatchNo (None, 8, 8, 2048) 8192 ['conv5_block3_3_conv[0][0]']
rmalization)
conv5_block3_add (Add) (None, 8, 8, 2048) 0 ['conv5_block2_out[0][0]',
'conv5_block3_3_bn[0][0]']
conv5_block3_out (Activati (None, 8, 8, 2048) 0 ['conv5_block3_add[0][0]']
on)
==================================================================================================
Total params: 23587712 (89.98 MB)
Trainable params: 23534592 (89.78 MB)
Non-trainable params: 53120 (207.50 KB)
__________________________________________________________________________________________________
# freeze the model
for layer in basemodel.layers:
layers.trainable = False
# add a classification head to the model
headmodel = basemodel.output
headmodel = AveragePooling2D(pool_size= (4,4))(headmodel)
headmodel = Flatten(name = 'flatten')(headmodel)
headmodel = Dense(256, activation= 'relu')(headmodel)
headmodel = Dropout(0.3)(headmodel)
headmodel = Dense(256, activation= 'relu')(headmodel)
headmodel = Dropout(0.3)(headmodel)
headmodel = Dense(2, activation= 'softmax')(headmodel)
model = Model(inputs = basemodel.input, outputs = headmodel)
model.summary()
Model: "model"
__________________________________________________________________________________________________
Layer (type) Output Shape Param # Connected to
==================================================================================================
input_1 (InputLayer) [(None, 256, 256, 3)] 0 []
conv1_pad (ZeroPadding2D) (None, 262, 262, 3) 0 ['input_1[0][0]']
conv1_conv (Conv2D) (None, 128, 128, 64) 9472 ['conv1_pad[0][0]']
conv1_bn (BatchNormalizati (None, 128, 128, 64) 256 ['conv1_conv[0][0]']
on)
conv1_relu (Activation) (None, 128, 128, 64) 0 ['conv1_bn[0][0]']
pool1_pad (ZeroPadding2D) (None, 130, 130, 64) 0 ['conv1_relu[0][0]']
pool1_pool (MaxPooling2D) (None, 64, 64, 64) 0 ['pool1_pad[0][0]']
conv2_block1_1_conv (Conv2 (None, 64, 64, 64) 4160 ['pool1_pool[0][0]']
D)
conv2_block1_1_bn (BatchNo (None, 64, 64, 64) 256 ['conv2_block1_1_conv[0][0]']
rmalization)
conv2_block1_1_relu (Activ (None, 64, 64, 64) 0 ['conv2_block1_1_bn[0][0]']
ation)
conv2_block1_2_conv (Conv2 (None, 64, 64, 64) 36928 ['conv2_block1_1_relu[0][0]']
D)
conv2_block1_2_bn (BatchNo (None, 64, 64, 64) 256 ['conv2_block1_2_conv[0][0]']
rmalization)
conv2_block1_2_relu (Activ (None, 64, 64, 64) 0 ['conv2_block1_2_bn[0][0]']
ation)
conv2_block1_0_conv (Conv2 (None, 64, 64, 256) 16640 ['pool1_pool[0][0]']
D)
conv2_block1_3_conv (Conv2 (None, 64, 64, 256) 16640 ['conv2_block1_2_relu[0][0]']
D)
conv2_block1_0_bn (BatchNo (None, 64, 64, 256) 1024 ['conv2_block1_0_conv[0][0]']
rmalization)
conv2_block1_3_bn (BatchNo (None, 64, 64, 256) 1024 ['conv2_block1_3_conv[0][0]']
rmalization)
conv2_block1_add (Add) (None, 64, 64, 256) 0 ['conv2_block1_0_bn[0][0]',
'conv2_block1_3_bn[0][0]']
conv2_block1_out (Activati (None, 64, 64, 256) 0 ['conv2_block1_add[0][0]']
on)
conv2_block2_1_conv (Conv2 (None, 64, 64, 64) 16448 ['conv2_block1_out[0][0]']
D)
conv2_block2_1_bn (BatchNo (None, 64, 64, 64) 256 ['conv2_block2_1_conv[0][0]']
rmalization)
conv2_block2_1_relu (Activ (None, 64, 64, 64) 0 ['conv2_block2_1_bn[0][0]']
ation)
conv2_block2_2_conv (Conv2 (None, 64, 64, 64) 36928 ['conv2_block2_1_relu[0][0]']
D)
conv2_block2_2_bn (BatchNo (None, 64, 64, 64) 256 ['conv2_block2_2_conv[0][0]']
rmalization)
conv2_block2_2_relu (Activ (None, 64, 64, 64) 0 ['conv2_block2_2_bn[0][0]']
ation)
conv2_block2_3_conv (Conv2 (None, 64, 64, 256) 16640 ['conv2_block2_2_relu[0][0]']
D)
conv2_block2_3_bn (BatchNo (None, 64, 64, 256) 1024 ['conv2_block2_3_conv[0][0]']
rmalization)
conv2_block2_add (Add) (None, 64, 64, 256) 0 ['conv2_block1_out[0][0]',
'conv2_block2_3_bn[0][0]']
conv2_block2_out (Activati (None, 64, 64, 256) 0 ['conv2_block2_add[0][0]']
on)
conv2_block3_1_conv (Conv2 (None, 64, 64, 64) 16448 ['conv2_block2_out[0][0]']
D)
conv2_block3_1_bn (BatchNo (None, 64, 64, 64) 256 ['conv2_block3_1_conv[0][0]']
rmalization)
conv2_block3_1_relu (Activ (None, 64, 64, 64) 0 ['conv2_block3_1_bn[0][0]']
ation)
conv2_block3_2_conv (Conv2 (None, 64, 64, 64) 36928 ['conv2_block3_1_relu[0][0]']
D)
conv2_block3_2_bn (BatchNo (None, 64, 64, 64) 256 ['conv2_block3_2_conv[0][0]']
rmalization)
conv2_block3_2_relu (Activ (None, 64, 64, 64) 0 ['conv2_block3_2_bn[0][0]']
ation)
conv2_block3_3_conv (Conv2 (None, 64, 64, 256) 16640 ['conv2_block3_2_relu[0][0]']
D)
conv2_block3_3_bn (BatchNo (None, 64, 64, 256) 1024 ['conv2_block3_3_conv[0][0]']
rmalization)
conv2_block3_add (Add) (None, 64, 64, 256) 0 ['conv2_block2_out[0][0]',
'conv2_block3_3_bn[0][0]']
conv2_block3_out (Activati (None, 64, 64, 256) 0 ['conv2_block3_add[0][0]']
on)
conv3_block1_1_conv (Conv2 (None, 32, 32, 128) 32896 ['conv2_block3_out[0][0]']
D)
conv3_block1_1_bn (BatchNo (None, 32, 32, 128) 512 ['conv3_block1_1_conv[0][0]']
rmalization)
conv3_block1_1_relu (Activ (None, 32, 32, 128) 0 ['conv3_block1_1_bn[0][0]']
ation)
conv3_block1_2_conv (Conv2 (None, 32, 32, 128) 147584 ['conv3_block1_1_relu[0][0]']
D)
conv3_block1_2_bn (BatchNo (None, 32, 32, 128) 512 ['conv3_block1_2_conv[0][0]']
rmalization)
conv3_block1_2_relu (Activ (None, 32, 32, 128) 0 ['conv3_block1_2_bn[0][0]']
ation)
conv3_block1_0_conv (Conv2 (None, 32, 32, 512) 131584 ['conv2_block3_out[0][0]']
D)
conv3_block1_3_conv (Conv2 (None, 32, 32, 512) 66048 ['conv3_block1_2_relu[0][0]']
D)
conv3_block1_0_bn (BatchNo (None, 32, 32, 512) 2048 ['conv3_block1_0_conv[0][0]']
rmalization)
conv3_block1_3_bn (BatchNo (None, 32, 32, 512) 2048 ['conv3_block1_3_conv[0][0]']
rmalization)
conv3_block1_add (Add) (None, 32, 32, 512) 0 ['conv3_block1_0_bn[0][0]',
'conv3_block1_3_bn[0][0]']
conv3_block1_out (Activati (None, 32, 32, 512) 0 ['conv3_block1_add[0][0]']
on)
conv3_block2_1_conv (Conv2 (None, 32, 32, 128) 65664 ['conv3_block1_out[0][0]']
D)
conv3_block2_1_bn (BatchNo (None, 32, 32, 128) 512 ['conv3_block2_1_conv[0][0]']
rmalization)
conv3_block2_1_relu (Activ (None, 32, 32, 128) 0 ['conv3_block2_1_bn[0][0]']
ation)
conv3_block2_2_conv (Conv2 (None, 32, 32, 128) 147584 ['conv3_block2_1_relu[0][0]']
D)
conv3_block2_2_bn (BatchNo (None, 32, 32, 128) 512 ['conv3_block2_2_conv[0][0]']
rmalization)
conv3_block2_2_relu (Activ (None, 32, 32, 128) 0 ['conv3_block2_2_bn[0][0]']
ation)
conv3_block2_3_conv (Conv2 (None, 32, 32, 512) 66048 ['conv3_block2_2_relu[0][0]']
D)
conv3_block2_3_bn (BatchNo (None, 32, 32, 512) 2048 ['conv3_block2_3_conv[0][0]']
rmalization)
conv3_block2_add (Add) (None, 32, 32, 512) 0 ['conv3_block1_out[0][0]',
'conv3_block2_3_bn[0][0]']
conv3_block2_out (Activati (None, 32, 32, 512) 0 ['conv3_block2_add[0][0]']
on)
conv3_block3_1_conv (Conv2 (None, 32, 32, 128) 65664 ['conv3_block2_out[0][0]']
D)
conv3_block3_1_bn (BatchNo (None, 32, 32, 128) 512 ['conv3_block3_1_conv[0][0]']
rmalization)
conv3_block3_1_relu (Activ (None, 32, 32, 128) 0 ['conv3_block3_1_bn[0][0]']
ation)
conv3_block3_2_conv (Conv2 (None, 32, 32, 128) 147584 ['conv3_block3_1_relu[0][0]']
D)
conv3_block3_2_bn (BatchNo (None, 32, 32, 128) 512 ['conv3_block3_2_conv[0][0]']
rmalization)
conv3_block3_2_relu (Activ (None, 32, 32, 128) 0 ['conv3_block3_2_bn[0][0]']
ation)
conv3_block3_3_conv (Conv2 (None, 32, 32, 512) 66048 ['conv3_block3_2_relu[0][0]']
D)
conv3_block3_3_bn (BatchNo (None, 32, 32, 512) 2048 ['conv3_block3_3_conv[0][0]']
rmalization)
conv3_block3_add (Add) (None, 32, 32, 512) 0 ['conv3_block2_out[0][0]',
'conv3_block3_3_bn[0][0]']
conv3_block3_out (Activati (None, 32, 32, 512) 0 ['conv3_block3_add[0][0]']
on)
conv3_block4_1_conv (Conv2 (None, 32, 32, 128) 65664 ['conv3_block3_out[0][0]']
D)
conv3_block4_1_bn (BatchNo (None, 32, 32, 128) 512 ['conv3_block4_1_conv[0][0]']
rmalization)
conv3_block4_1_relu (Activ (None, 32, 32, 128) 0 ['conv3_block4_1_bn[0][0]']
ation)
conv3_block4_2_conv (Conv2 (None, 32, 32, 128) 147584 ['conv3_block4_1_relu[0][0]']
D)
conv3_block4_2_bn (BatchNo (None, 32, 32, 128) 512 ['conv3_block4_2_conv[0][0]']
rmalization)
conv3_block4_2_relu (Activ (None, 32, 32, 128) 0 ['conv3_block4_2_bn[0][0]']
ation)
conv3_block4_3_conv (Conv2 (None, 32, 32, 512) 66048 ['conv3_block4_2_relu[0][0]']
D)
conv3_block4_3_bn (BatchNo (None, 32, 32, 512) 2048 ['conv3_block4_3_conv[0][0]']
rmalization)
conv3_block4_add (Add) (None, 32, 32, 512) 0 ['conv3_block3_out[0][0]',
'conv3_block4_3_bn[0][0]']
conv3_block4_out (Activati (None, 32, 32, 512) 0 ['conv3_block4_add[0][0]']
on)
conv4_block1_1_conv (Conv2 (None, 16, 16, 256) 131328 ['conv3_block4_out[0][0]']
D)
conv4_block1_1_bn (BatchNo (None, 16, 16, 256) 1024 ['conv4_block1_1_conv[0][0]']
rmalization)
conv4_block1_1_relu (Activ (None, 16, 16, 256) 0 ['conv4_block1_1_bn[0][0]']
ation)
conv4_block1_2_conv (Conv2 (None, 16, 16, 256) 590080 ['conv4_block1_1_relu[0][0]']
D)
conv4_block1_2_bn (BatchNo (None, 16, 16, 256) 1024 ['conv4_block1_2_conv[0][0]']
rmalization)
conv4_block1_2_relu (Activ (None, 16, 16, 256) 0 ['conv4_block1_2_bn[0][0]']
ation)
conv4_block1_0_conv (Conv2 (None, 16, 16, 1024) 525312 ['conv3_block4_out[0][0]']
D)
conv4_block1_3_conv (Conv2 (None, 16, 16, 1024) 263168 ['conv4_block1_2_relu[0][0]']
D)
conv4_block1_0_bn (BatchNo (None, 16, 16, 1024) 4096 ['conv4_block1_0_conv[0][0]']
rmalization)
conv4_block1_3_bn (BatchNo (None, 16, 16, 1024) 4096 ['conv4_block1_3_conv[0][0]']
rmalization)
conv4_block1_add (Add) (None, 16, 16, 1024) 0 ['conv4_block1_0_bn[0][0]',
'conv4_block1_3_bn[0][0]']
conv4_block1_out (Activati (None, 16, 16, 1024) 0 ['conv4_block1_add[0][0]']
on)
conv4_block2_1_conv (Conv2 (None, 16, 16, 256) 262400 ['conv4_block1_out[0][0]']
D)
conv4_block2_1_bn (BatchNo (None, 16, 16, 256) 1024 ['conv4_block2_1_conv[0][0]']
rmalization)
conv4_block2_1_relu (Activ (None, 16, 16, 256) 0 ['conv4_block2_1_bn[0][0]']
ation)
conv4_block2_2_conv (Conv2 (None, 16, 16, 256) 590080 ['conv4_block2_1_relu[0][0]']
D)
conv4_block2_2_bn (BatchNo (None, 16, 16, 256) 1024 ['conv4_block2_2_conv[0][0]']
rmalization)
conv4_block2_2_relu (Activ (None, 16, 16, 256) 0 ['conv4_block2_2_bn[0][0]']
ation)
conv4_block2_3_conv (Conv2 (None, 16, 16, 1024) 263168 ['conv4_block2_2_relu[0][0]']
D)
conv4_block2_3_bn (BatchNo (None, 16, 16, 1024) 4096 ['conv4_block2_3_conv[0][0]']
rmalization)
conv4_block2_add (Add) (None, 16, 16, 1024) 0 ['conv4_block1_out[0][0]',
'conv4_block2_3_bn[0][0]']
conv4_block2_out (Activati (None, 16, 16, 1024) 0 ['conv4_block2_add[0][0]']
on)
conv4_block3_1_conv (Conv2 (None, 16, 16, 256) 262400 ['conv4_block2_out[0][0]']
D)
conv4_block3_1_bn (BatchNo (None, 16, 16, 256) 1024 ['conv4_block3_1_conv[0][0]']
rmalization)
conv4_block3_1_relu (Activ (None, 16, 16, 256) 0 ['conv4_block3_1_bn[0][0]']
ation)
conv4_block3_2_conv (Conv2 (None, 16, 16, 256) 590080 ['conv4_block3_1_relu[0][0]']
D)
conv4_block3_2_bn (BatchNo (None, 16, 16, 256) 1024 ['conv4_block3_2_conv[0][0]']
rmalization)
conv4_block3_2_relu (Activ (None, 16, 16, 256) 0 ['conv4_block3_2_bn[0][0]']
ation)
conv4_block3_3_conv (Conv2 (None, 16, 16, 1024) 263168 ['conv4_block3_2_relu[0][0]']
D)
conv4_block3_3_bn (BatchNo (None, 16, 16, 1024) 4096 ['conv4_block3_3_conv[0][0]']
rmalization)
conv4_block3_add (Add) (None, 16, 16, 1024) 0 ['conv4_block2_out[0][0]',
'conv4_block3_3_bn[0][0]']
conv4_block3_out (Activati (None, 16, 16, 1024) 0 ['conv4_block3_add[0][0]']
on)
conv4_block4_1_conv (Conv2 (None, 16, 16, 256) 262400 ['conv4_block3_out[0][0]']
D)
conv4_block4_1_bn (BatchNo (None, 16, 16, 256) 1024 ['conv4_block4_1_conv[0][0]']
rmalization)
conv4_block4_1_relu (Activ (None, 16, 16, 256) 0 ['conv4_block4_1_bn[0][0]']
ation)
conv4_block4_2_conv (Conv2 (None, 16, 16, 256) 590080 ['conv4_block4_1_relu[0][0]']
D)
conv4_block4_2_bn (BatchNo (None, 16, 16, 256) 1024 ['conv4_block4_2_conv[0][0]']
rmalization)
conv4_block4_2_relu (Activ (None, 16, 16, 256) 0 ['conv4_block4_2_bn[0][0]']
ation)
conv4_block4_3_conv (Conv2 (None, 16, 16, 1024) 263168 ['conv4_block4_2_relu[0][0]']
D)
conv4_block4_3_bn (BatchNo (None, 16, 16, 1024) 4096 ['conv4_block4_3_conv[0][0]']
rmalization)
conv4_block4_add (Add) (None, 16, 16, 1024) 0 ['conv4_block3_out[0][0]',
'conv4_block4_3_bn[0][0]']
conv4_block4_out (Activati (None, 16, 16, 1024) 0 ['conv4_block4_add[0][0]']
on)
conv4_block5_1_conv (Conv2 (None, 16, 16, 256) 262400 ['conv4_block4_out[0][0]']
D)
conv4_block5_1_bn (BatchNo (None, 16, 16, 256) 1024 ['conv4_block5_1_conv[0][0]']
rmalization)
conv4_block5_1_relu (Activ (None, 16, 16, 256) 0 ['conv4_block5_1_bn[0][0]']
ation)
conv4_block5_2_conv (Conv2 (None, 16, 16, 256) 590080 ['conv4_block5_1_relu[0][0]']
D)
conv4_block5_2_bn (BatchNo (None, 16, 16, 256) 1024 ['conv4_block5_2_conv[0][0]']
rmalization)
conv4_block5_2_relu (Activ (None, 16, 16, 256) 0 ['conv4_block5_2_bn[0][0]']
ation)
conv4_block5_3_conv (Conv2 (None, 16, 16, 1024) 263168 ['conv4_block5_2_relu[0][0]']
D)
conv4_block5_3_bn (BatchNo (None, 16, 16, 1024) 4096 ['conv4_block5_3_conv[0][0]']
rmalization)
conv4_block5_add (Add) (None, 16, 16, 1024) 0 ['conv4_block4_out[0][0]',
'conv4_block5_3_bn[0][0]']
conv4_block5_out (Activati (None, 16, 16, 1024) 0 ['conv4_block5_add[0][0]']
on)
conv4_block6_1_conv (Conv2 (None, 16, 16, 256) 262400 ['conv4_block5_out[0][0]']
D)
conv4_block6_1_bn (BatchNo (None, 16, 16, 256) 1024 ['conv4_block6_1_conv[0][0]']
rmalization)
conv4_block6_1_relu (Activ (None, 16, 16, 256) 0 ['conv4_block6_1_bn[0][0]']
ation)
conv4_block6_2_conv (Conv2 (None, 16, 16, 256) 590080 ['conv4_block6_1_relu[0][0]']
D)
conv4_block6_2_bn (BatchNo (None, 16, 16, 256) 1024 ['conv4_block6_2_conv[0][0]']
rmalization)
conv4_block6_2_relu (Activ (None, 16, 16, 256) 0 ['conv4_block6_2_bn[0][0]']
ation)
conv4_block6_3_conv (Conv2 (None, 16, 16, 1024) 263168 ['conv4_block6_2_relu[0][0]']
D)
conv4_block6_3_bn (BatchNo (None, 16, 16, 1024) 4096 ['conv4_block6_3_conv[0][0]']
rmalization)
conv4_block6_add (Add) (None, 16, 16, 1024) 0 ['conv4_block5_out[0][0]',
'conv4_block6_3_bn[0][0]']
conv4_block6_out (Activati (None, 16, 16, 1024) 0 ['conv4_block6_add[0][0]']
on)
conv5_block1_1_conv (Conv2 (None, 8, 8, 512) 524800 ['conv4_block6_out[0][0]']
D)
conv5_block1_1_bn (BatchNo (None, 8, 8, 512) 2048 ['conv5_block1_1_conv[0][0]']
rmalization)
conv5_block1_1_relu (Activ (None, 8, 8, 512) 0 ['conv5_block1_1_bn[0][0]']
ation)
conv5_block1_2_conv (Conv2 (None, 8, 8, 512) 2359808 ['conv5_block1_1_relu[0][0]']
D)
conv5_block1_2_bn (BatchNo (None, 8, 8, 512) 2048 ['conv5_block1_2_conv[0][0]']
rmalization)
conv5_block1_2_relu (Activ (None, 8, 8, 512) 0 ['conv5_block1_2_bn[0][0]']
ation)
conv5_block1_0_conv (Conv2 (None, 8, 8, 2048) 2099200 ['conv4_block6_out[0][0]']
D)
conv5_block1_3_conv (Conv2 (None, 8, 8, 2048) 1050624 ['conv5_block1_2_relu[0][0]']
D)
conv5_block1_0_bn (BatchNo (None, 8, 8, 2048) 8192 ['conv5_block1_0_conv[0][0]']
rmalization)
conv5_block1_3_bn (BatchNo (None, 8, 8, 2048) 8192 ['conv5_block1_3_conv[0][0]']
rmalization)
conv5_block1_add (Add) (None, 8, 8, 2048) 0 ['conv5_block1_0_bn[0][0]',
'conv5_block1_3_bn[0][0]']
conv5_block1_out (Activati (None, 8, 8, 2048) 0 ['conv5_block1_add[0][0]']
on)
conv5_block2_1_conv (Conv2 (None, 8, 8, 512) 1049088 ['conv5_block1_out[0][0]']
D)
conv5_block2_1_bn (BatchNo (None, 8, 8, 512) 2048 ['conv5_block2_1_conv[0][0]']
rmalization)
conv5_block2_1_relu (Activ (None, 8, 8, 512) 0 ['conv5_block2_1_bn[0][0]']
ation)
conv5_block2_2_conv (Conv2 (None, 8, 8, 512) 2359808 ['conv5_block2_1_relu[0][0]']
D)
conv5_block2_2_bn (BatchNo (None, 8, 8, 512) 2048 ['conv5_block2_2_conv[0][0]']
rmalization)
conv5_block2_2_relu (Activ (None, 8, 8, 512) 0 ['conv5_block2_2_bn[0][0]']
ation)
conv5_block2_3_conv (Conv2 (None, 8, 8, 2048) 1050624 ['conv5_block2_2_relu[0][0]']
D)
conv5_block2_3_bn (BatchNo (None, 8, 8, 2048) 8192 ['conv5_block2_3_conv[0][0]']
rmalization)
conv5_block2_add (Add) (None, 8, 8, 2048) 0 ['conv5_block1_out[0][0]',
'conv5_block2_3_bn[0][0]']
conv5_block2_out (Activati (None, 8, 8, 2048) 0 ['conv5_block2_add[0][0]']
on)
conv5_block3_1_conv (Conv2 (None, 8, 8, 512) 1049088 ['conv5_block2_out[0][0]']
D)
conv5_block3_1_bn (BatchNo (None, 8, 8, 512) 2048 ['conv5_block3_1_conv[0][0]']
rmalization)
conv5_block3_1_relu (Activ (None, 8, 8, 512) 0 ['conv5_block3_1_bn[0][0]']
ation)
conv5_block3_2_conv (Conv2 (None, 8, 8, 512) 2359808 ['conv5_block3_1_relu[0][0]']
D)
conv5_block3_2_bn (BatchNo (None, 8, 8, 512) 2048 ['conv5_block3_2_conv[0][0]']
rmalization)
conv5_block3_2_relu (Activ (None, 8, 8, 512) 0 ['conv5_block3_2_bn[0][0]']
ation)
conv5_block3_3_conv (Conv2 (None, 8, 8, 2048) 1050624 ['conv5_block3_2_relu[0][0]']
D)
conv5_block3_3_bn (BatchNo (None, 8, 8, 2048) 8192 ['conv5_block3_3_conv[0][0]']
rmalization)
conv5_block3_add (Add) (None, 8, 8, 2048) 0 ['conv5_block2_out[0][0]',
'conv5_block3_3_bn[0][0]']
conv5_block3_out (Activati (None, 8, 8, 2048) 0 ['conv5_block3_add[0][0]']
on)
average_pooling2d (Average (None, 2, 2, 2048) 0 ['conv5_block3_out[0][0]']
Pooling2D)
flatten (Flatten) (None, 8192) 0 ['average_pooling2d[0][0]']
dense (Dense) (None, 256) 2097408 ['flatten[0][0]']
dropout (Dropout) (None, 256) 0 ['dense[0][0]']
dense_1 (Dense) (None, 256) 65792 ['dropout[0][0]']
dropout_1 (Dropout) (None, 256) 0 ['dense_1[0][0]']
dense_2 (Dense) (None, 2) 514 ['dropout_1[0][0]']
==================================================================================================
Total params: 25751426 (98.23 MB)
Trainable params: 25698306 (98.03 MB)
Non-trainable params: 53120 (207.50 KB)
__________________________________________________________________________________________________
# compile the model
model.compile(loss= 'categorical_crossentropy', optimizer = 'adam', metrics = ['accuracy'])
WARNING:tensorflow:From C:\Users\olayi\anaconda3\Lib\site-packages\keras\src\optimizers\__init__.py:309: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.
# use early stopping to exit training if validation loss is not decreasing even after certain epochs (patience)
earlystopping = EarlyStopping(monitor = 'val_loss', mode = 'min', verbose = 1, patience = 20)
# save the best model with the least validation loss
checkpointer = ModelCheckpoint(filepath='classifier-resnet-weights.hdf5', verbose = 1, save_best_only = True)
history = model.fit(train_generator, steps_per_epoch=train_generator.n // 16, epochs = 10, validation_data= val_generator, validation_steps= val_generator.n // 16, callbacks= [checkpointer, earlystopping])
Epoch 1/10 WARNING:tensorflow:From C:\Users\olayi\anaconda3\Lib\site-packages\keras\src\utils\tf_utils.py:492: The name tf.ragged.RaggedTensorValue is deprecated. Please use tf.compat.v1.ragged.RaggedTensorValue instead. WARNING:tensorflow:From C:\Users\olayi\anaconda3\Lib\site-packages\keras\src\engine\base_layer_utils.py:384: The name tf.executing_eagerly_outside_functions is deprecated. Please use tf.compat.v1.executing_eagerly_outside_functions instead. 177/177 [==============================] - ETA: 0s - loss: 0.9508 - accuracy: 0.6532 Epoch 1: val_loss improved from inf to 0.65904, saving model to classifier-resnet-weights.hdf5 177/177 [==============================] - 755s 4s/step - loss: 0.9508 - accuracy: 0.6532 - val_loss: 0.6590 - val_accuracy: 0.6331 Epoch 2/10 177/177 [==============================] - ETA: 0s - loss: 0.5868 - accuracy: 0.7535 Epoch 2: val_loss did not improve from 0.65904 177/177 [==============================] - 693s 4s/step - loss: 0.5868 - accuracy: 0.7535 - val_loss: 0.6681 - val_accuracy: 0.6331 Epoch 3/10 177/177 [==============================] - ETA: 0s - loss: 0.4516 - accuracy: 0.7977 Epoch 3: val_loss did not improve from 0.65904 177/177 [==============================] - 663s 4s/step - loss: 0.4516 - accuracy: 0.7977 - val_loss: 0.6843 - val_accuracy: 0.6351 Epoch 4/10 177/177 [==============================] - ETA: 0s - loss: 0.3710 - accuracy: 0.8353 Epoch 4: val_loss did not improve from 0.65904 177/177 [==============================] - 705s 4s/step - loss: 0.3710 - accuracy: 0.8353 - val_loss: 0.7307 - val_accuracy: 0.6310 Epoch 5/10 177/177 [==============================] - ETA: 0s - loss: 0.2947 - accuracy: 0.8700 Epoch 5: val_loss did not improve from 0.65904 177/177 [==============================] - 5530s 31s/step - loss: 0.2947 - accuracy: 0.8700 - val_loss: 0.6770 - val_accuracy: 0.6694 Epoch 6/10 177/177 [==============================] - ETA: 0s - loss: 0.2807 - accuracy: 0.8874 Epoch 6: val_loss improved from 0.65904 to 0.51411, saving model to classifier-resnet-weights.hdf5 177/177 [==============================] - 1010s 6s/step - loss: 0.2807 - accuracy: 0.8874 - val_loss: 0.5141 - val_accuracy: 0.7722 Epoch 7/10 177/177 [==============================] - ETA: 0s - loss: 0.3441 - accuracy: 0.8753 Epoch 7: val_loss did not improve from 0.51411 177/177 [==============================] - 141604s 805s/step - loss: 0.3441 - accuracy: 0.8753 - val_loss: 0.7713 - val_accuracy: 0.6976 Epoch 8/10 177/177 [==============================] - ETA: 0s - loss: 0.2707 - accuracy: 0.8916 Epoch 8: val_loss improved from 0.51411 to 0.31863, saving model to classifier-resnet-weights.hdf5 177/177 [==============================] - 701s 4s/step - loss: 0.2707 - accuracy: 0.8916 - val_loss: 0.3186 - val_accuracy: 0.8609 Epoch 9/10 177/177 [==============================] - ETA: 0s - loss: 0.2450 - accuracy: 0.9065 Epoch 9: val_loss did not improve from 0.31863 177/177 [==============================] - 11596s 66s/step - loss: 0.2450 - accuracy: 0.9065 - val_loss: 0.3296 - val_accuracy: 0.8589 Epoch 10/10 177/177 [==============================] - ETA: 0s - loss: 0.2187 - accuracy: 0.9175 Epoch 10: val_loss did not improve from 0.31863 177/177 [==============================] - 8441s 48s/step - loss: 0.2187 - accuracy: 0.9175 - val_loss: 0.3216 - val_accuracy: 0.8891
# save model architecture to json file for future use
model_json = model.to_json()
with open('classifier_resnet_model.json', 'w') as json_file:
json_file.write(model_json)
# load json_file
with open('classifier_resnet_model.json', 'r') as json_file:
json_savedModel = json_file.read()
model = tf.keras.models.model_from_json(json_savedModel)
model.load_weights('classifier-resnet-weights.hdf5')
model.compile(loss = 'categorical_crossentropy', optimizer= 'adam', metrics = ['accuracy'])
# make prediction
test_pred = model.predict(test_generator, steps = test_generator.n // 16, verbose = 1)
36/36 [==============================] - 38s 956ms/step
test_pred.shape
(576, 2)
test_pred
array([[9.9453026e-01, 5.4697674e-03],
[9.7976160e-01, 2.0238433e-02],
[4.4421285e-01, 5.5578715e-01],
...,
[9.9669302e-01, 3.3069481e-03],
[9.2080944e-12, 1.0000000e+00],
[9.5426548e-01, 4.5734555e-02]], dtype=float32)
# obtain the predicted class from the model prediction
predict = []
for i in test_pred:
predict.append(str(np.argmax(i)))
predict = np.asarray(predict)
predict
array(['0', '0', '1', '1', '1', '0', '0', '1', '1', '0', '0', '0', '0',
'0', '0', '1', '1', '1', '0', '1', '1', '1', '0', '0', '0', '1',
'0', '1', '0', '1', '1', '0', '0', '0', '1', '0', '0', '0', '0',
'0', '0', '1', '1', '1', '0', '1', '1', '0', '1', '0', '1', '0',
'1', '0', '0', '0', '1', '1', '0', '1', '0', '0', '0', '0', '1',
'0', '1', '0', '0', '1', '1', '0', '0', '1', '0', '1', '0', '1',
'1', '0', '0', '0', '0', '0', '0', '0', '0', '0', '1', '0', '1',
'0', '0', '0', '0', '0', '0', '1', '0', '0', '1', '1', '0', '0',
'0', '1', '0', '1', '0', '0', '0', '0', '1', '0', '1', '1', '1',
'1', '1', '1', '0', '0', '0', '0', '1', '1', '0', '1', '1', '1',
'0', '0', '1', '1', '1', '0', '0', '0', '1', '1', '1', '1', '0',
'0', '0', '0', '1', '0', '1', '0', '0', '0', '1', '1', '1', '0',
'1', '0', '0', '0', '1', '1', '0', '0', '1', '1', '0', '0', '0',
'0', '0', '1', '1', '1', '0', '0', '1', '0', '1', '0', '0', '1',
'0', '1', '1', '0', '1', '1', '1', '1', '0', '1', '1', '1', '0',
'0', '0', '1', '1', '0', '0', '0', '1', '1', '1', '1', '0', '1',
'1', '1', '0', '0', '0', '1', '0', '0', '0', '1', '0', '0', '1',
'1', '0', '0', '1', '0', '1', '1', '1', '0', '1', '1', '0', '0',
'0', '0', '0', '0', '0', '0', '1', '0', '0', '0', '0', '0', '1',
'1', '1', '1', '0', '0', '0', '0', '0', '0', '0', '1', '1', '1',
'0', '1', '0', '0', '0', '1', '0', '0', '1', '1', '1', '1', '0',
'0', '1', '0', '0', '0', '0', '0', '0', '0', '1', '0', '1', '0',
'0', '1', '1', '1', '1', '1', '0', '0', '0', '0', '0', '1', '1',
'0', '0', '1', '0', '0', '1', '0', '1', '1', '0', '0', '0', '0',
'1', '0', '0', '1', '0', '0', '1', '0', '0', '0', '0', '1', '1',
'0', '1', '1', '0', '0', '0', '1', '0', '1', '0', '0', '0', '1',
'1', '0', '0', '0', '1', '0', '1', '0', '1', '0', '0', '1', '1',
'0', '1', '1', '0', '1', '1', '0', '1', '0', '0', '1', '0', '0',
'0', '1', '1', '0', '1', '1', '0', '1', '0', '0', '1', '0', '1',
'0', '1', '0', '1', '1', '0', '0', '0', '1', '0', '1', '0', '0',
'0', '1', '0', '1', '0', '0', '1', '0', '1', '0', '0', '1', '0',
'0', '1', '0', '0', '0', '1', '0', '0', '1', '0', '0', '0', '1',
'1', '1', '0', '0', '0', '0', '1', '0', '0', '0', '1', '0', '0',
'0', '1', '0', '1', '0', '1', '1', '0', '0', '0', '1', '0', '0',
'0', '1', '0', '1', '1', '1', '0', '0', '0', '1', '1', '1', '1',
'0', '1', '1', '0', '1', '0', '0', '0', '0', '0', '1', '0', '0',
'0', '0', '0', '1', '0', '0', '0', '0', '0', '0', '0', '1', '0',
'0', '0', '0', '0', '0', '1', '1', '1', '0', '0', '0', '0', '1',
'1', '1', '0', '0', '0', '1', '1', '0', '1', '0', '0', '0', '0',
'0', '1', '0', '0', '1', '0', '0', '0', '1', '0', '1', '1', '0',
'1', '0', '0', '1', '1', '1', '0', '1', '1', '1', '0', '0', '1',
'0', '0', '0', '0', '0', '0', '0', '0', '1', '0', '0', '0', '0',
'0', '1', '1', '0', '0', '1', '1', '0', '0', '0', '0', '0', '0',
'0', '0', '1', '0', '0', '1', '0', '0', '0', '1', '1', '1', '0',
'1', '0', '1', '0'], dtype='<U1')
# compare the predictions to the original test data
# since we have used test generator, it limited the images to len(predict), due to batch size
original_test = np.asarray(test['mask'])[:len(predict)]
len(original_test)
576
# obtain the accuracy of the model
from sklearn.metrics import accuracy_score
accuracy = accuracy_score(original_test, predict)
accuracy
0.8697916666666666
# plot the confusion matrix
from sklearn.metrics import confusion_matrix
cm = confusion_matrix(original_test, predict)
plt.figure(figsize = (7,7))
sns.heatmap(cm, annot = True)
<Axes: >
# Obatin the classification report
from sklearn.metrics import classification_report
report = classification_report(original_test, predict, labels = [0,1])
print(report)
precision recall f1-score support
0 0.94 0.85 0.90 378
1 0.76 0.90 0.83 198
micro avg 0.87 0.87 0.87 576
macro avg 0.85 0.88 0.86 576
weighted avg 0.88 0.87 0.87 576
# Get the dataframe containing MRIs which have masks associated with them
brain_df_mask = brain_df[brain_df['mask']==1]
brain_df_mask.shape
(1373, 4)
# split the data into train the test set
from sklearn.model_selection import train_test_split
X_train, X_val = train_test_split(brain_df_mask, test_size = 0.15)
X_test, X_val = train_test_split(X_val, test_size = 0.5)
print('X_train ', X_train)
print('X_val ', X_val)
print('X_test ', X_test)
X_train patient_id \
2744 TCGA_HT_7855_19951020
3358 TCGA_HT_A5RC_19990831
1773 TCGA_DU_A5TT_19980318
3016 TCGA_HT_7879_19981009
2691 TCGA_HT_7856_19950831
... ...
1684 TCGA_DU_8165_19970205
2746 TCGA_HT_7855_19951020
2986 TCGA_HT_7882_19970125
3480 TCGA_DU_8164_19970111
3532 TCGA_DU_8168_19970503
image_path \
2744 TCGA_FG_6690_20020226/TCGA_FG_6690_20020226_27...
3358 TCGA_DU_6401_19831001/TCGA_DU_6401_19831001_37...
1773 TCGA_CS_6665_20010817/TCGA_CS_6665_20010817_17...
3016 TCGA_FG_6690_20020226/TCGA_FG_6690_20020226_31...
2691 TCGA_HT_A61A_20000127/TCGA_HT_A61A_20000127_26...
... ...
1684 TCGA_DU_7008_19830723/TCGA_DU_7008_19830723_16...
2746 TCGA_FG_7643_20021104/TCGA_FG_7643_20021104_27...
2986 TCGA_DU_6399_19830416/TCGA_DU_6399_19830416_31...
3480 TCGA_DU_6405_19851005/TCGA_DU_6405_19851005_42...
3532 TCGA_DU_7014_19860618/TCGA_DU_7014_19860618_44...
mask_path mask
2744 TCGA_FG_6690_20020226/TCGA_FG_6690_20020226_27... 1
3358 TCGA_DU_6401_19831001/TCGA_DU_6401_19831001_37... 1
1773 TCGA_CS_6665_20010817/TCGA_CS_6665_20010817_17... 1
3016 TCGA_FG_6690_20020226/TCGA_FG_6690_20020226_31... 1
2691 TCGA_HT_A61A_20000127/TCGA_HT_A61A_20000127_26... 1
... ... ...
1684 TCGA_DU_7008_19830723/TCGA_DU_7008_19830723_16... 1
2746 TCGA_FG_7643_20021104/TCGA_FG_7643_20021104_27... 1
2986 TCGA_DU_6399_19830416/TCGA_DU_6399_19830416_31... 1
3480 TCGA_DU_6405_19851005/TCGA_DU_6405_19851005_42... 1
3532 TCGA_DU_7014_19860618/TCGA_DU_7014_19860618_44... 1
[1167 rows x 4 columns]
X_val patient_id \
2667 TCGA_HT_7616_19940813
1601 TCGA_DU_A5TP_19970614
2150 TCGA_FG_6691_20020405
2009 TCGA_FG_6689_20020326
1519 TCGA_DU_7306_19930512
... ...
3042 TCGA_HT_7877_19980917
3205 TCGA_HT_8563_19981209
2239 TCGA_FG_7634_20000128
3390 TCGA_HT_A61B_19991127
1835 TCGA_FG_5962_20000626
image_path \
2667 TCGA_DU_A5TY_19970709/TCGA_DU_A5TY_19970709_26...
1601 TCGA_FG_7637_20000922/TCGA_FG_7637_20000922_15...
2150 TCGA_FG_7634_20000128/TCGA_FG_7634_20000128_20...
2009 TCGA_DU_7018_19911220/TCGA_DU_7018_19911220_19...
1519 TCGA_HT_8113_19930809/TCGA_HT_8113_19930809_14...
... ...
3042 TCGA_DU_6401_19831001/TCGA_DU_6401_19831001_31...
3205 TCGA_HT_A61A_20000127/TCGA_HT_A61A_20000127_34...
2239 TCGA_DU_8166_19970322/TCGA_DU_8166_19970322_21...
3390 TCGA_DU_7010_19860307/TCGA_DU_7010_19860307_38...
1835 TCGA_HT_7692_19960724/TCGA_HT_7692_19960724_17...
mask_path mask
2667 TCGA_DU_A5TY_19970709/TCGA_DU_A5TY_19970709_26... 1
1601 TCGA_FG_7637_20000922/TCGA_FG_7637_20000922_15... 1
2150 TCGA_FG_7634_20000128/TCGA_FG_7634_20000128_20... 1
2009 TCGA_DU_7018_19911220/TCGA_DU_7018_19911220_19... 1
1519 TCGA_HT_8113_19930809/TCGA_HT_8113_19930809_14... 1
... ... ...
3042 TCGA_DU_6401_19831001/TCGA_DU_6401_19831001_31... 1
3205 TCGA_HT_A61A_20000127/TCGA_HT_A61A_20000127_34... 1
2239 TCGA_DU_8166_19970322/TCGA_DU_8166_19970322_21... 1
3390 TCGA_DU_7010_19860307/TCGA_DU_7010_19860307_38... 1
1835 TCGA_HT_7692_19960724/TCGA_HT_7692_19960724_17... 1
[103 rows x 4 columns]
X_test patient_id \
1816 TCGA_DU_A5TY_19970709
3783 TCGA_DU_7304_19930325
1484 TCGA_DU_7306_19930512
1842 TCGA_FG_5962_20000626
995 TCGA_DU_6405_19851005
... ...
1794 TCGA_DU_A5TT_19980318
1639 TCGA_DU_8166_19970322
2771 TCGA_HT_7692_19960724
2154 TCGA_FG_6691_20020405
1286 TCGA_DU_8163_19961119
image_path \
1816 TCGA_FG_6692_20020606/TCGA_FG_6692_20020606_17...
3783 TCGA_FG_A60K_20040224/TCGA_FG_A60K_20040224_56...
1484 TCGA_DU_A5TU_19980312/TCGA_DU_A5TU_19980312_14...
1842 TCGA_HT_7882_19970125/TCGA_HT_7882_19970125_17...
995 TCGA_CS_5393_19990606/TCGA_CS_5393_19990606_10...
... ...
1794 TCGA_DU_7008_19830723/TCGA_DU_7008_19830723_17...
1639 TCGA_DU_8168_19970503/TCGA_DU_8168_19970503_15...
2771 TCGA_FG_6688_20020215/TCGA_FG_6688_20020215_27...
2154 TCGA_FG_A4MU_20030903/TCGA_FG_A4MU_20030903_20...
1286 TCGA_HT_7874_19950902/TCGA_HT_7874_19950902_12...
mask_path mask
1816 TCGA_FG_6692_20020606/TCGA_FG_6692_20020606_17... 1
3783 TCGA_FG_A60K_20040224/TCGA_FG_A60K_20040224_56... 1
1484 TCGA_DU_A5TU_19980312/TCGA_DU_A5TU_19980312_14... 1
1842 TCGA_HT_7882_19970125/TCGA_HT_7882_19970125_17... 1
995 TCGA_CS_5393_19990606/TCGA_CS_5393_19990606_10... 1
... ... ...
1794 TCGA_DU_7008_19830723/TCGA_DU_7008_19830723_17... 1
1639 TCGA_DU_8168_19970503/TCGA_DU_8168_19970503_15... 1
2771 TCGA_FG_6688_20020215/TCGA_FG_6688_20020215_27... 1
2154 TCGA_FG_A4MU_20030903/TCGA_FG_A4MU_20030903_20... 1
1286 TCGA_HT_7874_19950902/TCGA_HT_7874_19950902_12... 1
[103 rows x 4 columns]
# Create a seperate list for the imageId, classId to pass into the generator
train_ids = list(X_train.image_path)
train_mask = list(X_train.mask_path)
val_ids = list(X_val.image_path)
val_mask = list(X_val.mask_path)
# utilities file contains the code for custom loass function and custom data generator
from utilities import DataGenerator
training_gen = DataGenerator(train_ids, train_mask)
validation_gen = DataGenerator(val_ids, val_mask)
# building ResUNet model
# Using the res block
def resblock(X, f):
# make a copy of the input
X_copy = X
# main path
X = Conv2D(f, kernel_size = (1,1), strides = (1,1), kernel_initializer= 'he_normal')(X)
X = BatchNormalization()(X)
X = Activation('relu')(X)
X = Conv2D(f, kernel_size = (3,3), strides = (1,1), padding = 'same', kernel_initializer= 'he_normal')(X)
X = BatchNormalization()(X)
# Short path
X_copy = Conv2D(f, kernel_size = (1,1), strides = (1,1), kernel_initializer= 'he_normal')(X_copy)
X_copy = BatchNormalization()(X_copy)
# Adding the output from the main path and the short path together
X = add([X,X_copy])
X = Activation('relu')(X)
return X
# function to upscale and concetenate the values passed
def upsample_concat(x, skip):
x = UpSampling2D((2,2))(x)
merge = Concatenate()([x, skip])
return merge
# build the network
input_shape = (256, 256, 3)
# Input tensor shape
X_input = Input(input_shape)
# Stage 1
conv1_in = Conv2D(16,3, activation ='relu', padding = 'same', kernel_initializer = 'he_normal')(X_input)
conv1_in = BatchNormalization()(conv1_in)
conv1_in = Conv2D(16,3, activation ='relu', padding = 'same', kernel_initializer = 'he_normal')(conv1_in)
conv1_in = BatchNormalization()(conv1_in)
pool_1 = MaxPool2D(pool_size= (2,2))(conv1_in)
# Stage 2
conv2_in = resblock(pool_1,32)
pool_2 = MaxPool2D(pool_size= (2,2))(conv2_in)
# Stage 3
conv3_in = resblock(pool_2,64)
pool_3 = MaxPool2D(pool_size= (2,2))(conv3_in)
# Stage 4
conv4_in = resblock(pool_3,128)
pool_4 = MaxPool2D(pool_size= (2,2))(conv4_in)
# Stage 5 (Bottle Neck)
conv5_in = resblock(pool_4,256)
# Upsale stage 1
up_1 = upsample_concat(conv5_in, conv4_in)
up_1 = resblock(up_1, 128)
# Upsale stage 2
up_2 = upsample_concat(up_1, conv3_in)
up_2 = resblock(up_2, 64)
# Upsale stage 3
up_3 = upsample_concat(up_2, conv2_in)
up_3 = resblock(up_3, 32)
# Upsale stage 4
up_4 = upsample_concat(up_3, conv1_in)
up_4 = resblock(up_4, 16)
# Final Output
output = Conv2D(1, (1,1), padding = 'same', activation = 'sigmoid')(up_4)
model_seg = Model(inputs = X_input, outputs = output)
model_seg.summary()
Model: "model_1"
__________________________________________________________________________________________________
Layer (type) Output Shape Param # Connected to
==================================================================================================
input_2 (InputLayer) [(None, 256, 256, 3)] 0 []
conv2d (Conv2D) (None, 256, 256, 16) 448 ['input_2[0][0]']
batch_normalization (Batch (None, 256, 256, 16) 64 ['conv2d[0][0]']
Normalization)
conv2d_1 (Conv2D) (None, 256, 256, 16) 2320 ['batch_normalization[0][0]']
batch_normalization_1 (Bat (None, 256, 256, 16) 64 ['conv2d_1[0][0]']
chNormalization)
max_pooling2d (MaxPooling2 (None, 128, 128, 16) 0 ['batch_normalization_1[0][0]'
D) ]
conv2d_2 (Conv2D) (None, 128, 128, 32) 544 ['max_pooling2d[0][0]']
batch_normalization_2 (Bat (None, 128, 128, 32) 128 ['conv2d_2[0][0]']
chNormalization)
activation (Activation) (None, 128, 128, 32) 0 ['batch_normalization_2[0][0]'
]
conv2d_3 (Conv2D) (None, 128, 128, 32) 9248 ['activation[0][0]']
conv2d_4 (Conv2D) (None, 128, 128, 32) 544 ['max_pooling2d[0][0]']
batch_normalization_3 (Bat (None, 128, 128, 32) 128 ['conv2d_3[0][0]']
chNormalization)
batch_normalization_4 (Bat (None, 128, 128, 32) 128 ['conv2d_4[0][0]']
chNormalization)
add (Add) (None, 128, 128, 32) 0 ['batch_normalization_3[0][0]'
, 'batch_normalization_4[0][0]
']
activation_1 (Activation) (None, 128, 128, 32) 0 ['add[0][0]']
max_pooling2d_1 (MaxPoolin (None, 64, 64, 32) 0 ['activation_1[0][0]']
g2D)
conv2d_5 (Conv2D) (None, 64, 64, 64) 2112 ['max_pooling2d_1[0][0]']
batch_normalization_5 (Bat (None, 64, 64, 64) 256 ['conv2d_5[0][0]']
chNormalization)
activation_2 (Activation) (None, 64, 64, 64) 0 ['batch_normalization_5[0][0]'
]
conv2d_6 (Conv2D) (None, 64, 64, 64) 36928 ['activation_2[0][0]']
conv2d_7 (Conv2D) (None, 64, 64, 64) 2112 ['max_pooling2d_1[0][0]']
batch_normalization_6 (Bat (None, 64, 64, 64) 256 ['conv2d_6[0][0]']
chNormalization)
batch_normalization_7 (Bat (None, 64, 64, 64) 256 ['conv2d_7[0][0]']
chNormalization)
add_1 (Add) (None, 64, 64, 64) 0 ['batch_normalization_6[0][0]'
, 'batch_normalization_7[0][0]
']
activation_3 (Activation) (None, 64, 64, 64) 0 ['add_1[0][0]']
max_pooling2d_2 (MaxPoolin (None, 32, 32, 64) 0 ['activation_3[0][0]']
g2D)
conv2d_8 (Conv2D) (None, 32, 32, 128) 8320 ['max_pooling2d_2[0][0]']
batch_normalization_8 (Bat (None, 32, 32, 128) 512 ['conv2d_8[0][0]']
chNormalization)
activation_4 (Activation) (None, 32, 32, 128) 0 ['batch_normalization_8[0][0]'
]
conv2d_9 (Conv2D) (None, 32, 32, 128) 147584 ['activation_4[0][0]']
conv2d_10 (Conv2D) (None, 32, 32, 128) 8320 ['max_pooling2d_2[0][0]']
batch_normalization_9 (Bat (None, 32, 32, 128) 512 ['conv2d_9[0][0]']
chNormalization)
batch_normalization_10 (Ba (None, 32, 32, 128) 512 ['conv2d_10[0][0]']
tchNormalization)
add_2 (Add) (None, 32, 32, 128) 0 ['batch_normalization_9[0][0]'
, 'batch_normalization_10[0][0
]']
activation_5 (Activation) (None, 32, 32, 128) 0 ['add_2[0][0]']
max_pooling2d_3 (MaxPoolin (None, 16, 16, 128) 0 ['activation_5[0][0]']
g2D)
conv2d_11 (Conv2D) (None, 16, 16, 256) 33024 ['max_pooling2d_3[0][0]']
batch_normalization_11 (Ba (None, 16, 16, 256) 1024 ['conv2d_11[0][0]']
tchNormalization)
activation_6 (Activation) (None, 16, 16, 256) 0 ['batch_normalization_11[0][0]
']
conv2d_12 (Conv2D) (None, 16, 16, 256) 590080 ['activation_6[0][0]']
conv2d_13 (Conv2D) (None, 16, 16, 256) 33024 ['max_pooling2d_3[0][0]']
batch_normalization_12 (Ba (None, 16, 16, 256) 1024 ['conv2d_12[0][0]']
tchNormalization)
batch_normalization_13 (Ba (None, 16, 16, 256) 1024 ['conv2d_13[0][0]']
tchNormalization)
add_3 (Add) (None, 16, 16, 256) 0 ['batch_normalization_12[0][0]
',
'batch_normalization_13[0][0]
']
activation_7 (Activation) (None, 16, 16, 256) 0 ['add_3[0][0]']
up_sampling2d (UpSampling2 (None, 32, 32, 256) 0 ['activation_7[0][0]']
D)
concatenate (Concatenate) (None, 32, 32, 384) 0 ['up_sampling2d[0][0]',
'activation_5[0][0]']
conv2d_14 (Conv2D) (None, 32, 32, 128) 49280 ['concatenate[0][0]']
batch_normalization_14 (Ba (None, 32, 32, 128) 512 ['conv2d_14[0][0]']
tchNormalization)
activation_8 (Activation) (None, 32, 32, 128) 0 ['batch_normalization_14[0][0]
']
conv2d_15 (Conv2D) (None, 32, 32, 128) 147584 ['activation_8[0][0]']
conv2d_16 (Conv2D) (None, 32, 32, 128) 49280 ['concatenate[0][0]']
batch_normalization_15 (Ba (None, 32, 32, 128) 512 ['conv2d_15[0][0]']
tchNormalization)
batch_normalization_16 (Ba (None, 32, 32, 128) 512 ['conv2d_16[0][0]']
tchNormalization)
add_4 (Add) (None, 32, 32, 128) 0 ['batch_normalization_15[0][0]
',
'batch_normalization_16[0][0]
']
activation_9 (Activation) (None, 32, 32, 128) 0 ['add_4[0][0]']
up_sampling2d_1 (UpSamplin (None, 64, 64, 128) 0 ['activation_9[0][0]']
g2D)
concatenate_1 (Concatenate (None, 64, 64, 192) 0 ['up_sampling2d_1[0][0]',
) 'activation_3[0][0]']
conv2d_17 (Conv2D) (None, 64, 64, 64) 12352 ['concatenate_1[0][0]']
batch_normalization_17 (Ba (None, 64, 64, 64) 256 ['conv2d_17[0][0]']
tchNormalization)
activation_10 (Activation) (None, 64, 64, 64) 0 ['batch_normalization_17[0][0]
']
conv2d_18 (Conv2D) (None, 64, 64, 64) 36928 ['activation_10[0][0]']
conv2d_19 (Conv2D) (None, 64, 64, 64) 12352 ['concatenate_1[0][0]']
batch_normalization_18 (Ba (None, 64, 64, 64) 256 ['conv2d_18[0][0]']
tchNormalization)
batch_normalization_19 (Ba (None, 64, 64, 64) 256 ['conv2d_19[0][0]']
tchNormalization)
add_5 (Add) (None, 64, 64, 64) 0 ['batch_normalization_18[0][0]
',
'batch_normalization_19[0][0]
']
activation_11 (Activation) (None, 64, 64, 64) 0 ['add_5[0][0]']
up_sampling2d_2 (UpSamplin (None, 128, 128, 64) 0 ['activation_11[0][0]']
g2D)
concatenate_2 (Concatenate (None, 128, 128, 96) 0 ['up_sampling2d_2[0][0]',
) 'activation_1[0][0]']
conv2d_20 (Conv2D) (None, 128, 128, 32) 3104 ['concatenate_2[0][0]']
batch_normalization_20 (Ba (None, 128, 128, 32) 128 ['conv2d_20[0][0]']
tchNormalization)
activation_12 (Activation) (None, 128, 128, 32) 0 ['batch_normalization_20[0][0]
']
conv2d_21 (Conv2D) (None, 128, 128, 32) 9248 ['activation_12[0][0]']
conv2d_22 (Conv2D) (None, 128, 128, 32) 3104 ['concatenate_2[0][0]']
batch_normalization_21 (Ba (None, 128, 128, 32) 128 ['conv2d_21[0][0]']
tchNormalization)
batch_normalization_22 (Ba (None, 128, 128, 32) 128 ['conv2d_22[0][0]']
tchNormalization)
add_6 (Add) (None, 128, 128, 32) 0 ['batch_normalization_21[0][0]
',
'batch_normalization_22[0][0]
']
activation_13 (Activation) (None, 128, 128, 32) 0 ['add_6[0][0]']
up_sampling2d_3 (UpSamplin (None, 256, 256, 32) 0 ['activation_13[0][0]']
g2D)
concatenate_3 (Concatenate (None, 256, 256, 48) 0 ['up_sampling2d_3[0][0]',
) 'batch_normalization_1[0][0]'
]
conv2d_23 (Conv2D) (None, 256, 256, 16) 784 ['concatenate_3[0][0]']
batch_normalization_23 (Ba (None, 256, 256, 16) 64 ['conv2d_23[0][0]']
tchNormalization)
activation_14 (Activation) (None, 256, 256, 16) 0 ['batch_normalization_23[0][0]
']
conv2d_24 (Conv2D) (None, 256, 256, 16) 2320 ['activation_14[0][0]']
conv2d_25 (Conv2D) (None, 256, 256, 16) 784 ['concatenate_3[0][0]']
batch_normalization_24 (Ba (None, 256, 256, 16) 64 ['conv2d_24[0][0]']
tchNormalization)
batch_normalization_25 (Ba (None, 256, 256, 16) 64 ['conv2d_25[0][0]']
tchNormalization)
add_7 (Add) (None, 256, 256, 16) 0 ['batch_normalization_24[0][0]
',
'batch_normalization_25[0][0]
']
activation_15 (Activation) (None, 256, 256, 16) 0 ['add_7[0][0]']
conv2d_26 (Conv2D) (None, 256, 256, 1) 17 ['activation_15[0][0]']
==================================================================================================
Total params: 1210513 (4.62 MB)
Trainable params: 1206129 (4.60 MB)
Non-trainable params: 4384 (17.12 KB)
__________________________________________________________________________________________________
# using the custom loss function Focal_tversky, tversky_loss, tversky
from utilities import focal_tversky, tversky_loss, tversky
# code for focal_tversky
from keras.losses import binary_crossentropy
import keras.backend as K
import tensorflow as tf
epsilon = 1e-5
smooth = 1
def dsc(y_true, y_pred):
smooth = 1.
y_true_f = K.flatten(y_true)
y_pred_f = K.flatten(y_pred)
intersection = K.sum(y_true_f * y_pred_f)
score = (2. * intersection + smooth) / (K.sum(y_true_f) + K.sum(y_pred_f) + smooth)
return score
def dice_loss(y_true, y_pred):
loss = 1 - dsc(y_true, y_pred)
return loss
def bce_dice_loss(y_true, y_pred):
loss = binary_crossentropy(y_true, y_pred) + dice_loss(y_true, y_pred)
return loss
def confusion(y_true, y_pred):
smooth=1
y_pred_pos = K.clip(y_pred, 0, 1)
y_pred_neg = 1 - y_pred_pos
y_pos = K.clip(y_true, 0, 1)
y_neg = 1 - y_pos
tp = K.sum(y_pos * y_pred_pos)
fp = K.sum(y_neg * y_pred_pos)
fn = K.sum(y_pos * y_pred_neg)
prec = (tp + smooth)/(tp+fp+smooth)
recall = (tp+smooth)/(tp+fn+smooth)
return prec, recall
def tp(y_true, y_pred):
smooth = 1
y_pred_pos = K.round(K.clip(y_pred, 0, 1))
y_pos = K.round(K.clip(y_true, 0, 1))
tp = (K.sum(y_pos * y_pred_pos) + smooth)/ (K.sum(y_pos) + smooth)
return tp
def tn(y_true, y_pred):
smooth = 1
y_pred_pos = K.round(K.clip(y_pred, 0, 1))
y_pred_neg = 1 - y_pred_pos
y_pos = K.round(K.clip(y_true, 0, 1))
y_neg = 1 - y_pos
tn = (K.sum(y_neg * y_pred_neg) + smooth) / (K.sum(y_neg) + smooth )
return tn
def tversky(y_true, y_pred):
y_true_pos = K.flatten(y_true)
y_pred_pos = K.flatten(y_pred)
true_pos = K.sum(y_true_pos * y_pred_pos)
false_neg = K.sum(y_true_pos * (1-y_pred_pos))
false_pos = K.sum((1-y_true_pos)*y_pred_pos)
alpha = 0.7
return (true_pos + smooth)/(true_pos + alpha*false_neg + (1-alpha)*false_pos + smooth)
def tversky_loss(y_true, y_pred):
return 1 - tversky(y_true,y_pred)
def focal_tversky(y_true,y_pred):
y_true = tf.cast(y_true, tf.float32) # Cast y_true to float32
y_pred = tf.cast(y_pred, tf.float32)
pt_1 = tversky(y_true, y_pred)
gamma = 0.75
return K.pow((1-pt_1), gamma)
# compile model
adam = tf.keras.optimizers.legacy.Adam(lr = 0.05, epsilon = 0.1)
model_seg.compile(optimizer= adam, loss = focal_tversky, metrics = [tversky])
# use early stopping to exit training if validation loss is not decreasing even after certain epochs (patience)
earlystopping = EarlyStopping(monitor = 'val_loss', mode = 'min', verbose = 1, patience = 20)
# save the best model with the least validation loss
checkpointer = ModelCheckpoint(filepath='ResUNet-weights.hdf5', verbose = 1, save_best_only = True)
history = model_seg.fit(training_gen, epochs = 40, validation_data= validation_gen, callbacks= [checkpointer, earlystopping])
Epoch 1/40 72/72 [==============================] - ETA: 0s - loss: 0.8447 - tversky: 0.2007 Epoch 1: val_loss improved from inf to 0.75451, saving model to ResUNet-weights.hdf5 72/72 [==============================] - 102s 1s/step - loss: 0.8447 - tversky: 0.2007 - val_loss: 0.7545 - val_tversky: 0.3127 Epoch 2/40 72/72 [==============================] - ETA: 0s - loss: 0.5082 - tversky: 0.5898 Epoch 2: val_loss improved from 0.75451 to 0.47842, saving model to ResUNet-weights.hdf5 72/72 [==============================] - 96s 1s/step - loss: 0.5082 - tversky: 0.5898 - val_loss: 0.4784 - val_tversky: 0.6244 Epoch 3/40 72/72 [==============================] - ETA: 0s - loss: 0.3789 - tversky: 0.7236 Epoch 3: val_loss did not improve from 0.47842 72/72 [==============================] - 106s 1s/step - loss: 0.3789 - tversky: 0.7236 - val_loss: 0.5118 - val_tversky: 0.5892 Epoch 4/40 72/72 [==============================] - ETA: 0s - loss: 0.3349 - tversky: 0.7658 Epoch 4: val_loss did not improve from 0.47842 72/72 [==============================] - 108s 2s/step - loss: 0.3349 - tversky: 0.7658 - val_loss: 0.5491 - val_tversky: 0.5476 Epoch 5/40 72/72 [==============================] - ETA: 0s - loss: 0.3042 - tversky: 0.7933 Epoch 5: val_loss improved from 0.47842 to 0.34035, saving model to ResUNet-weights.hdf5 72/72 [==============================] - 149s 2s/step - loss: 0.3042 - tversky: 0.7933 - val_loss: 0.3403 - val_tversky: 0.7612 Epoch 6/40 72/72 [==============================] - ETA: 0s - loss: 0.2718 - tversky: 0.8224 Epoch 6: val_loss improved from 0.34035 to 0.28135, saving model to ResUNet-weights.hdf5 72/72 [==============================] - 7448s 105s/step - loss: 0.2718 - tversky: 0.8224 - val_loss: 0.2813 - val_tversky: 0.8152 Epoch 7/40 72/72 [==============================] - ETA: 0s - loss: 0.2590 - tversky: 0.8336 Epoch 7: val_loss did not improve from 0.28135 72/72 [==============================] - 146s 2s/step - loss: 0.2590 - tversky: 0.8336 - val_loss: 0.4022 - val_tversky: 0.7023 Epoch 8/40 72/72 [==============================] - ETA: 0s - loss: 0.2463 - tversky: 0.8442 Epoch 8: val_loss did not improve from 0.28135 72/72 [==============================] - 141s 2s/step - loss: 0.2463 - tversky: 0.8442 - val_loss: 0.3328 - val_tversky: 0.7686 Epoch 9/40 72/72 [==============================] - ETA: 0s - loss: 0.2289 - tversky: 0.8588 Epoch 9: val_loss did not improve from 0.28135 72/72 [==============================] - 138s 2s/step - loss: 0.2289 - tversky: 0.8588 - val_loss: 0.3500 - val_tversky: 0.7532 Epoch 10/40 72/72 [==============================] - ETA: 0s - loss: 0.2254 - tversky: 0.8606 Epoch 10: val_loss improved from 0.28135 to 0.26285, saving model to ResUNet-weights.hdf5 72/72 [==============================] - 139s 2s/step - loss: 0.2254 - tversky: 0.8606 - val_loss: 0.2629 - val_tversky: 0.8311 Epoch 11/40 72/72 [==============================] - ETA: 0s - loss: 0.2121 - tversky: 0.8720 Epoch 11: val_loss improved from 0.26285 to 0.24320, saving model to ResUNet-weights.hdf5 72/72 [==============================] - 141s 2s/step - loss: 0.2121 - tversky: 0.8720 - val_loss: 0.2432 - val_tversky: 0.8473 Epoch 12/40 72/72 [==============================] - ETA: 0s - loss: 0.1986 - tversky: 0.8831 Epoch 12: val_loss did not improve from 0.24320 72/72 [==============================] - 134s 2s/step - loss: 0.1986 - tversky: 0.8831 - val_loss: 0.2750 - val_tversky: 0.8198 Epoch 13/40 72/72 [==============================] - ETA: 0s - loss: 0.2041 - tversky: 0.8785 Epoch 13: val_loss did not improve from 0.24320 72/72 [==============================] - 139s 2s/step - loss: 0.2041 - tversky: 0.8785 - val_loss: 0.2935 - val_tversky: 0.8044 Epoch 14/40 72/72 [==============================] - ETA: 0s - loss: 0.1905 - tversky: 0.8891 Epoch 14: val_loss did not improve from 0.24320 72/72 [==============================] - 131s 2s/step - loss: 0.1905 - tversky: 0.8891 - val_loss: 0.2896 - val_tversky: 0.8071 Epoch 15/40 72/72 [==============================] - ETA: 0s - loss: 0.1786 - tversky: 0.8986 Epoch 15: val_loss improved from 0.24320 to 0.21935, saving model to ResUNet-weights.hdf5 72/72 [==============================] - 133s 2s/step - loss: 0.1786 - tversky: 0.8986 - val_loss: 0.2194 - val_tversky: 0.8659 Epoch 16/40 72/72 [==============================] - ETA: 0s - loss: 0.1713 - tversky: 0.9041 Epoch 16: val_loss did not improve from 0.21935 72/72 [==============================] - 134s 2s/step - loss: 0.1713 - tversky: 0.9041 - val_loss: 0.2678 - val_tversky: 0.8269 Epoch 17/40 72/72 [==============================] - ETA: 0s - loss: 0.1593 - tversky: 0.9130 Epoch 17: val_loss improved from 0.21935 to 0.21474, saving model to ResUNet-weights.hdf5 72/72 [==============================] - 130s 2s/step - loss: 0.1593 - tversky: 0.9130 - val_loss: 0.2147 - val_tversky: 0.8710 Epoch 18/40 72/72 [==============================] - ETA: 0s - loss: 0.1494 - tversky: 0.9202 Epoch 18: val_loss improved from 0.21474 to 0.20514, saving model to ResUNet-weights.hdf5 72/72 [==============================] - 134s 2s/step - loss: 0.1494 - tversky: 0.9202 - val_loss: 0.2051 - val_tversky: 0.8777 Epoch 19/40 72/72 [==============================] - ETA: 0s - loss: 0.1488 - tversky: 0.9205 Epoch 19: val_loss did not improve from 0.20514 72/72 [==============================] - 133s 2s/step - loss: 0.1488 - tversky: 0.9205 - val_loss: 0.2670 - val_tversky: 0.8268 Epoch 20/40 72/72 [==============================] - ETA: 0s - loss: 0.1505 - tversky: 0.9193 Epoch 20: val_loss improved from 0.20514 to 0.19853, saving model to ResUNet-weights.hdf5 72/72 [==============================] - 133s 2s/step - loss: 0.1505 - tversky: 0.9193 - val_loss: 0.1985 - val_tversky: 0.8839 Epoch 21/40 72/72 [==============================] - ETA: 0s - loss: 0.1416 - tversky: 0.9258 Epoch 21: val_loss improved from 0.19853 to 0.18324, saving model to ResUNet-weights.hdf5 72/72 [==============================] - 133s 2s/step - loss: 0.1416 - tversky: 0.9258 - val_loss: 0.1832 - val_tversky: 0.8954 Epoch 22/40 72/72 [==============================] - ETA: 0s - loss: 0.1495 - tversky: 0.9200 Epoch 22: val_loss did not improve from 0.18324 72/72 [==============================] - 133s 2s/step - loss: 0.1495 - tversky: 0.9200 - val_loss: 0.1991 - val_tversky: 0.8829 Epoch 23/40 72/72 [==============================] - ETA: 0s - loss: 0.1456 - tversky: 0.9228 Epoch 23: val_loss did not improve from 0.18324 72/72 [==============================] - 132s 2s/step - loss: 0.1456 - tversky: 0.9228 - val_loss: 0.2055 - val_tversky: 0.8777 Epoch 24/40 72/72 [==============================] - ETA: 0s - loss: 0.1454 - tversky: 0.9231 Epoch 24: val_loss did not improve from 0.18324 72/72 [==============================] - 132s 2s/step - loss: 0.1454 - tversky: 0.9231 - val_loss: 0.2337 - val_tversky: 0.8553 Epoch 25/40 72/72 [==============================] - ETA: 0s - loss: 0.1339 - tversky: 0.9312 Epoch 25: val_loss did not improve from 0.18324 72/72 [==============================] - 133s 2s/step - loss: 0.1339 - tversky: 0.9312 - val_loss: 0.1893 - val_tversky: 0.8908 Epoch 26/40 72/72 [==============================] - ETA: 0s - loss: 0.1268 - tversky: 0.9361 Epoch 26: val_loss improved from 0.18324 to 0.17267, saving model to ResUNet-weights.hdf5 72/72 [==============================] - 670s 9s/step - loss: 0.1268 - tversky: 0.9361 - val_loss: 0.1727 - val_tversky: 0.9037 Epoch 27/40 72/72 [==============================] - ETA: 0s - loss: 0.1204 - tversky: 0.9403 Epoch 27: val_loss did not improve from 0.17267 72/72 [==============================] - 532s 7s/step - loss: 0.1204 - tversky: 0.9403 - val_loss: 0.1820 - val_tversky: 0.8968 Epoch 28/40 72/72 [==============================] - ETA: 0s - loss: 0.1201 - tversky: 0.9405 Epoch 28: val_loss did not improve from 0.17267 72/72 [==============================] - 134s 2s/step - loss: 0.1201 - tversky: 0.9405 - val_loss: 0.1813 - val_tversky: 0.8972 Epoch 29/40 72/72 [==============================] - ETA: 0s - loss: 0.1212 - tversky: 0.9397 Epoch 29: val_loss did not improve from 0.17267 72/72 [==============================] - 133s 2s/step - loss: 0.1212 - tversky: 0.9397 - val_loss: 0.1850 - val_tversky: 0.8944 Epoch 30/40 72/72 [==============================] - ETA: 0s - loss: 0.1179 - tversky: 0.9419 Epoch 30: val_loss did not improve from 0.17267 72/72 [==============================] - 134s 2s/step - loss: 0.1179 - tversky: 0.9419 - val_loss: 0.1781 - val_tversky: 0.8995 Epoch 31/40 72/72 [==============================] - ETA: 0s - loss: 0.1125 - tversky: 0.9455 Epoch 31: val_loss did not improve from 0.17267 72/72 [==============================] - 134s 2s/step - loss: 0.1125 - tversky: 0.9455 - val_loss: 0.1937 - val_tversky: 0.8875 Epoch 32/40 72/72 [==============================] - ETA: 0s - loss: 0.1082 - tversky: 0.9482 Epoch 32: val_loss improved from 0.17267 to 0.17189, saving model to ResUNet-weights.hdf5 72/72 [==============================] - 132s 2s/step - loss: 0.1082 - tversky: 0.9482 - val_loss: 0.1719 - val_tversky: 0.9043 Epoch 33/40 72/72 [==============================] - ETA: 0s - loss: 0.1077 - tversky: 0.9485 Epoch 33: val_loss did not improve from 0.17189 72/72 [==============================] - 132s 2s/step - loss: 0.1077 - tversky: 0.9485 - val_loss: 0.1899 - val_tversky: 0.8907 Epoch 34/40 72/72 [==============================] - ETA: 0s - loss: 0.1118 - tversky: 0.9459 Epoch 34: val_loss did not improve from 0.17189 72/72 [==============================] - 131s 2s/step - loss: 0.1118 - tversky: 0.9459 - val_loss: 0.1929 - val_tversky: 0.8874 Epoch 35/40 72/72 [==============================] - ETA: 0s - loss: 0.1065 - tversky: 0.9493 Epoch 35: val_loss did not improve from 0.17189 72/72 [==============================] - 132s 2s/step - loss: 0.1065 - tversky: 0.9493 - val_loss: 0.1810 - val_tversky: 0.8973 Epoch 36/40 72/72 [==============================] - ETA: 0s - loss: 0.1038 - tversky: 0.9510 Epoch 36: val_loss did not improve from 0.17189 72/72 [==============================] - 131s 2s/step - loss: 0.1038 - tversky: 0.9510 - val_loss: 0.2112 - val_tversky: 0.8725 Epoch 37/40 72/72 [==============================] - ETA: 0s - loss: 0.1012 - tversky: 0.9527 Epoch 37: val_loss did not improve from 0.17189 72/72 [==============================] - 131s 2s/step - loss: 0.1012 - tversky: 0.9527 - val_loss: 0.1852 - val_tversky: 0.8936 Epoch 38/40 72/72 [==============================] - ETA: 0s - loss: 0.0993 - tversky: 0.9539 Epoch 38: val_loss did not improve from 0.17189 72/72 [==============================] - 131s 2s/step - loss: 0.0993 - tversky: 0.9539 - val_loss: 0.1868 - val_tversky: 0.8928 Epoch 39/40 72/72 [==============================] - ETA: 0s - loss: 0.0980 - tversky: 0.9547 Epoch 39: val_loss improved from 0.17189 to 0.16629, saving model to ResUNet-weights.hdf5 72/72 [==============================] - 131s 2s/step - loss: 0.0980 - tversky: 0.9547 - val_loss: 0.1663 - val_tversky: 0.9081 Epoch 40/40 72/72 [==============================] - ETA: 0s - loss: 0.0967 - tversky: 0.9554 Epoch 40: val_loss did not improve from 0.16629 72/72 [==============================] - 131s 2s/step - loss: 0.0967 - tversky: 0.9554 - val_loss: 0.1769 - val_tversky: 0.9004
# save model architecture to json file for future use
model_json = model_seg.to_json()
with open('ResUNet_model.json', 'w') as json_file:
json_file.write(model_json)
from utilities import focal_tversky, tversky_loss, tversky
with open('ResUNet_model.json', 'r') as json_file:
json_savedModel= json_file.read()
# load the model architecture
model_seg = tf.keras.models.model_from_json(json_savedModel)
model_seg.load_weights('ResUNet-weights.hdf5')
adam = tf.keras.optimizers.Adam(lr = 0.05, epsilon = 0.1)
model_seg.compile(optimizer = adam, loss = focal_tversky, metrics = [tversky])
WARNING:absl:`lr` is deprecated in Keras optimizer, please use `learning_rate` or use the legacy optimizer, e.g.,tf.keras.optimizers.legacy.Adam.
# Utilities file contains the code for custom loss function and custom data generator
from utilities import prediction
# making prediction
image_id, mask, has_mask = prediction(test, model, model_seg)
1/1 [==============================] - 0s 211ms/step 1/1 [==============================] - 0s 230ms/step 1/1 [==============================] - 0s 204ms/step 1/1 [==============================] - 0s 105ms/step 1/1 [==============================] - 0s 217ms/step 1/1 [==============================] - 0s 109ms/step 1/1 [==============================] - 0s 191ms/step 1/1 [==============================] - 0s 99ms/step 1/1 [==============================] - 0s 224ms/step 1/1 [==============================] - 0s 213ms/step 1/1 [==============================] - 0s 221ms/step 1/1 [==============================] - 0s 116ms/step 1/1 [==============================] - 0s 212ms/step 1/1 [==============================] - 0s 100ms/step 1/1 [==============================] - 0s 210ms/step 1/1 [==============================] - 0s 224ms/step 1/1 [==============================] - 0s 200ms/step 1/1 [==============================] - 0s 211ms/step 1/1 [==============================] - 0s 234ms/step 1/1 [==============================] - 0s 181ms/step 1/1 [==============================] - 0s 166ms/step 1/1 [==============================] - 0s 105ms/step 1/1 [==============================] - 0s 205ms/step 1/1 [==============================] - 0s 109ms/step 1/1 [==============================] - 0s 198ms/step 1/1 [==============================] - 0s 101ms/step 1/1 [==============================] - 0s 201ms/step 1/1 [==============================] - 0s 185ms/step 1/1 [==============================] - 0s 100ms/step 1/1 [==============================] - 0s 221ms/step 1/1 [==============================] - 0s 101ms/step 1/1 [==============================] - 0s 191ms/step 1/1 [==============================] - 0s 117ms/step 1/1 [==============================] - 0s 223ms/step 1/1 [==============================] - 0s 166ms/step 1/1 [==============================] - 0s 204ms/step 1/1 [==============================] - 0s 213ms/step 1/1 [==============================] - 0s 118ms/step 1/1 [==============================] - 0s 209ms/step 1/1 [==============================] - 0s 198ms/step 1/1 [==============================] - 0s 99ms/step 1/1 [==============================] - 0s 175ms/step 1/1 [==============================] - 0s 220ms/step 1/1 [==============================] - 0s 112ms/step 1/1 [==============================] - 0s 221ms/step 1/1 [==============================] - 0s 97ms/step 1/1 [==============================] - 0s 236ms/step 1/1 [==============================] - 0s 226ms/step 1/1 [==============================] - 0s 223ms/step 1/1 [==============================] - 0s 227ms/step 1/1 [==============================] - 0s 92ms/step 1/1 [==============================] - 0s 202ms/step 1/1 [==============================] - 0s 207ms/step 1/1 [==============================] - 0s 231ms/step 1/1 [==============================] - 0s 214ms/step 1/1 [==============================] - 0s 197ms/step 1/1 [==============================] - 0s 189ms/step 1/1 [==============================] - 0s 199ms/step 1/1 [==============================] - 0s 114ms/step 1/1 [==============================] - 0s 233ms/step 1/1 [==============================] - 0s 107ms/step 1/1 [==============================] - 0s 234ms/step 1/1 [==============================] - 0s 115ms/step 1/1 [==============================] - 0s 226ms/step 1/1 [==============================] - 0s 200ms/step 1/1 [==============================] - 0s 122ms/step 1/1 [==============================] - 0s 217ms/step 1/1 [==============================] - 0s 102ms/step 1/1 [==============================] - 0s 187ms/step 1/1 [==============================] - 0s 201ms/step 1/1 [==============================] - 0s 103ms/step 1/1 [==============================] - 0s 214ms/step 1/1 [==============================] - 0s 192ms/step 1/1 [==============================] - 0s 91ms/step 1/1 [==============================] - 0s 214ms/step 1/1 [==============================] - 0s 229ms/step 1/1 [==============================] - 0s 109ms/step 1/1 [==============================] - 0s 243ms/step 1/1 [==============================] - 0s 217ms/step 1/1 [==============================] - 0s 229ms/step 1/1 [==============================] - 0s 189ms/step 1/1 [==============================] - 0s 104ms/step 1/1 [==============================] - 0s 218ms/step 1/1 [==============================] - 0s 107ms/step 1/1 [==============================] - 0s 180ms/step 1/1 [==============================] - 0s 194ms/step 1/1 [==============================] - 0s 105ms/step 1/1 [==============================] - 0s 168ms/step 1/1 [==============================] - 0s 202ms/step 1/1 [==============================] - 0s 219ms/step 1/1 [==============================] - 0s 214ms/step 1/1 [==============================] - 0s 205ms/step 1/1 [==============================] - 0s 105ms/step 1/1 [==============================] - 0s 240ms/step 1/1 [==============================] - 0s 219ms/step 1/1 [==============================] - 0s 115ms/step 1/1 [==============================] - 0s 231ms/step 1/1 [==============================] - 0s 219ms/step 1/1 [==============================] - 0s 226ms/step 1/1 [==============================] - 0s 114ms/step 1/1 [==============================] - 0s 200ms/step 1/1 [==============================] - 0s 98ms/step 1/1 [==============================] - 0s 183ms/step 1/1 [==============================] - 0s 191ms/step 1/1 [==============================] - 0s 223ms/step 1/1 [==============================] - 0s 116ms/step 1/1 [==============================] - 0s 213ms/step 1/1 [==============================] - 0s 234ms/step 1/1 [==============================] - 0s 99ms/step 1/1 [==============================] - 0s 209ms/step 1/1 [==============================] - 0s 200ms/step 1/1 [==============================] - 0s 112ms/step 1/1 [==============================] - 0s 225ms/step 1/1 [==============================] - 0s 109ms/step 1/1 [==============================] - 0s 185ms/step 1/1 [==============================] - 0s 219ms/step 1/1 [==============================] - 0s 225ms/step 1/1 [==============================] - 0s 224ms/step 1/1 [==============================] - 0s 201ms/step 1/1 [==============================] - 0s 223ms/step 1/1 [==============================] - 0s 201ms/step 1/1 [==============================] - 0s 188ms/step 1/1 [==============================] - 0s 204ms/step 1/1 [==============================] - 0s 233ms/step 1/1 [==============================] - 0s 112ms/step 1/1 [==============================] - 0s 207ms/step 1/1 [==============================] - 0s 209ms/step 1/1 [==============================] - 0s 115ms/step 1/1 [==============================] - 0s 209ms/step 1/1 [==============================] - 0s 208ms/step 1/1 [==============================] - 0s 211ms/step 1/1 [==============================] - 0s 207ms/step 1/1 [==============================] - 0s 210ms/step 1/1 [==============================] - 0s 213ms/step 1/1 [==============================] - 0s 217ms/step 1/1 [==============================] - 0s 114ms/step 1/1 [==============================] - 0s 229ms/step 1/1 [==============================] - 0s 219ms/step 1/1 [==============================] - 0s 231ms/step 1/1 [==============================] - 0s 117ms/step 1/1 [==============================] - 0s 232ms/step 1/1 [==============================] - 0s 113ms/step 1/1 [==============================] - 0s 223ms/step 1/1 [==============================] - 0s 232ms/step 1/1 [==============================] - 0s 204ms/step 1/1 [==============================] - 0s 210ms/step 1/1 [==============================] - 0s 114ms/step 1/1 [==============================] - 0s 221ms/step 1/1 [==============================] - 0s 217ms/step 1/1 [==============================] - 0s 113ms/step 1/1 [==============================] - 0s 234ms/step 1/1 [==============================] - 0s 226ms/step 1/1 [==============================] - 0s 218ms/step 1/1 [==============================] - 0s 194ms/step 1/1 [==============================] - 0s 144ms/step 1/1 [==============================] - 0s 115ms/step 1/1 [==============================] - 0s 216ms/step 1/1 [==============================] - 0s 212ms/step 1/1 [==============================] - 0s 110ms/step 1/1 [==============================] - 0s 203ms/step 1/1 [==============================] - 0s 90ms/step 1/1 [==============================] - 0s 200ms/step 1/1 [==============================] - 0s 114ms/step 1/1 [==============================] - 0s 209ms/step 1/1 [==============================] - 0s 101ms/step 1/1 [==============================] - 0s 200ms/step 1/1 [==============================] - 0s 108ms/step 1/1 [==============================] - 0s 222ms/step 1/1 [==============================] - 0s 111ms/step 1/1 [==============================] - 0s 209ms/step 1/1 [==============================] - 0s 223ms/step 1/1 [==============================] - 0s 209ms/step 1/1 [==============================] - 0s 200ms/step 1/1 [==============================] - 0s 230ms/step 1/1 [==============================] - 0s 113ms/step 1/1 [==============================] - 0s 206ms/step 1/1 [==============================] - 0s 113ms/step 1/1 [==============================] - 0s 222ms/step 1/1 [==============================] - 0s 221ms/step 1/1 [==============================] - 0s 107ms/step 1/1 [==============================] - 0s 215ms/step 1/1 [==============================] - 0s 105ms/step 1/1 [==============================] - 0s 208ms/step 1/1 [==============================] - 0s 105ms/step 1/1 [==============================] - 0s 221ms/step 1/1 [==============================] - 0s 205ms/step 1/1 [==============================] - 0s 198ms/step 1/1 [==============================] - 0s 114ms/step 1/1 [==============================] - 0s 213ms/step 1/1 [==============================] - 0s 102ms/step 1/1 [==============================] - 0s 214ms/step 1/1 [==============================] - 0s 112ms/step 1/1 [==============================] - 0s 210ms/step 1/1 [==============================] - 0s 188ms/step 1/1 [==============================] - 0s 215ms/step 1/1 [==============================] - 0s 226ms/step 1/1 [==============================] - 0s 116ms/step 1/1 [==============================] - 0s 214ms/step 1/1 [==============================] - 0s 108ms/step 1/1 [==============================] - 0s 197ms/step 1/1 [==============================] - 0s 95ms/step 1/1 [==============================] - 0s 208ms/step 1/1 [==============================] - 0s 114ms/step 1/1 [==============================] - 0s 217ms/step 1/1 [==============================] - 0s 219ms/step 1/1 [==============================] - 0s 213ms/step 1/1 [==============================] - 0s 215ms/step 1/1 [==============================] - 0s 219ms/step 1/1 [==============================] - 0s 116ms/step 1/1 [==============================] - 0s 226ms/step 1/1 [==============================] - 0s 199ms/step 1/1 [==============================] - 0s 117ms/step 1/1 [==============================] - 0s 221ms/step 1/1 [==============================] - 0s 220ms/step 1/1 [==============================] - 0s 210ms/step 1/1 [==============================] - 0s 236ms/step 1/1 [==============================] - 0s 96ms/step 1/1 [==============================] - 0s 199ms/step 1/1 [==============================] - 0s 98ms/step 1/1 [==============================] - 0s 189ms/step 1/1 [==============================] - 0s 113ms/step 1/1 [==============================] - 0s 187ms/step 1/1 [==============================] - 0s 213ms/step 1/1 [==============================] - 0s 105ms/step 1/1 [==============================] - 0s 209ms/step 1/1 [==============================] - 0s 204ms/step 1/1 [==============================] - 0s 225ms/step 1/1 [==============================] - 0s 219ms/step 1/1 [==============================] - 0s 115ms/step 1/1 [==============================] - 0s 218ms/step 1/1 [==============================] - 0s 102ms/step 1/1 [==============================] - 0s 205ms/step 1/1 [==============================] - 0s 207ms/step 1/1 [==============================] - 0s 199ms/step 1/1 [==============================] - 0s 96ms/step 1/1 [==============================] - 0s 194ms/step 1/1 [==============================] - 0s 113ms/step 1/1 [==============================] - 0s 215ms/step 1/1 [==============================] - 0s 195ms/step 1/1 [==============================] - 0s 208ms/step 1/1 [==============================] - 0s 218ms/step 1/1 [==============================] - 0s 213ms/step 1/1 [==============================] - 0s 214ms/step 1/1 [==============================] - 0s 109ms/step 1/1 [==============================] - 0s 205ms/step 1/1 [==============================] - 0s 98ms/step 1/1 [==============================] - 0s 213ms/step 1/1 [==============================] - 0s 112ms/step 1/1 [==============================] - 0s 203ms/step 1/1 [==============================] - 0s 176ms/step 1/1 [==============================] - 0s 196ms/step 1/1 [==============================] - 0s 91ms/step 1/1 [==============================] - 0s 209ms/step 1/1 [==============================] - 0s 192ms/step 1/1 [==============================] - 0s 100ms/step 1/1 [==============================] - 0s 209ms/step 1/1 [==============================] - 0s 197ms/step 1/1 [==============================] - 0s 217ms/step 1/1 [==============================] - 0s 109ms/step 1/1 [==============================] - 0s 208ms/step 1/1 [==============================] - 0s 212ms/step 1/1 [==============================] - 0s 113ms/step 1/1 [==============================] - 0s 214ms/step 1/1 [==============================] - 0s 86ms/step 1/1 [==============================] - 0s 210ms/step 1/1 [==============================] - 0s 211ms/step 1/1 [==============================] - 0s 110ms/step 1/1 [==============================] - 0s 200ms/step 1/1 [==============================] - 0s 106ms/step 1/1 [==============================] - 0s 208ms/step 1/1 [==============================] - 0s 117ms/step 1/1 [==============================] - 0s 211ms/step 1/1 [==============================] - 0s 108ms/step 1/1 [==============================] - 0s 219ms/step 1/1 [==============================] - 0s 194ms/step 1/1 [==============================] - 0s 113ms/step 1/1 [==============================] - 0s 226ms/step 1/1 [==============================] - 0s 108ms/step 1/1 [==============================] - 0s 222ms/step 1/1 [==============================] - 0s 110ms/step 1/1 [==============================] - 0s 209ms/step 1/1 [==============================] - 0s 210ms/step 1/1 [==============================] - 0s 215ms/step 1/1 [==============================] - 0s 201ms/step 1/1 [==============================] - 0s 94ms/step 1/1 [==============================] - 0s 208ms/step 1/1 [==============================] - 0s 106ms/step 1/1 [==============================] - 0s 212ms/step 1/1 [==============================] - 0s 192ms/step 1/1 [==============================] - 0s 221ms/step 1/1 [==============================] - 0s 208ms/step 1/1 [==============================] - 0s 93ms/step 1/1 [==============================] - 0s 213ms/step 1/1 [==============================] - 0s 111ms/step 1/1 [==============================] - 0s 214ms/step 1/1 [==============================] - 0s 104ms/step 1/1 [==============================] - 0s 204ms/step 1/1 [==============================] - 0s 97ms/step 1/1 [==============================] - 0s 214ms/step 1/1 [==============================] - 0s 208ms/step 1/1 [==============================] - 0s 104ms/step 1/1 [==============================] - 0s 212ms/step 1/1 [==============================] - 0s 108ms/step 1/1 [==============================] - 0s 207ms/step 1/1 [==============================] - 0s 112ms/step 1/1 [==============================] - 0s 207ms/step 1/1 [==============================] - 0s 200ms/step 1/1 [==============================] - 0s 211ms/step 1/1 [==============================] - 0s 206ms/step 1/1 [==============================] - 0s 109ms/step 1/1 [==============================] - 0s 205ms/step 1/1 [==============================] - 0s 216ms/step 1/1 [==============================] - 0s 209ms/step 1/1 [==============================] - 0s 219ms/step 1/1 [==============================] - 0s 116ms/step 1/1 [==============================] - 0s 220ms/step 1/1 [==============================] - 0s 207ms/step 1/1 [==============================] - 0s 195ms/step 1/1 [==============================] - 0s 118ms/step 1/1 [==============================] - 0s 215ms/step 1/1 [==============================] - 0s 99ms/step 1/1 [==============================] - 0s 204ms/step 1/1 [==============================] - 0s 198ms/step 1/1 [==============================] - 0s 202ms/step 1/1 [==============================] - 0s 99ms/step 1/1 [==============================] - 0s 207ms/step 1/1 [==============================] - 0s 221ms/step 1/1 [==============================] - 0s 108ms/step 1/1 [==============================] - 0s 230ms/step 1/1 [==============================] - 0s 108ms/step 1/1 [==============================] - 0s 207ms/step 1/1 [==============================] - 0s 107ms/step 1/1 [==============================] - 0s 206ms/step 1/1 [==============================] - 0s 216ms/step 1/1 [==============================] - 0s 88ms/step 1/1 [==============================] - 0s 202ms/step 1/1 [==============================] - 0s 111ms/step 1/1 [==============================] - 0s 202ms/step 1/1 [==============================] - 0s 206ms/step 1/1 [==============================] - 0s 215ms/step 1/1 [==============================] - 0s 218ms/step 1/1 [==============================] - 0s 182ms/step 1/1 [==============================] - 0s 202ms/step 1/1 [==============================] - 0s 203ms/step 1/1 [==============================] - 0s 196ms/step 1/1 [==============================] - 0s 216ms/step 1/1 [==============================] - 0s 109ms/step 1/1 [==============================] - 0s 195ms/step 1/1 [==============================] - 0s 220ms/step 1/1 [==============================] - 0s 226ms/step 1/1 [==============================] - 0s 210ms/step 1/1 [==============================] - 0s 246ms/step 1/1 [==============================] - 0s 204ms/step 1/1 [==============================] - 0s 90ms/step 1/1 [==============================] - 0s 204ms/step 1/1 [==============================] - 0s 104ms/step 1/1 [==============================] - 0s 231ms/step 1/1 [==============================] - 0s 133ms/step 1/1 [==============================] - 0s 240ms/step 1/1 [==============================] - 0s 120ms/step 1/1 [==============================] - 0s 223ms/step 1/1 [==============================] - 0s 244ms/step 1/1 [==============================] - 0s 241ms/step 1/1 [==============================] - 0s 207ms/step 1/1 [==============================] - 0s 219ms/step 1/1 [==============================] - 0s 251ms/step 1/1 [==============================] - 0s 235ms/step 1/1 [==============================] - 0s 227ms/step 1/1 [==============================] - 0s 107ms/step 1/1 [==============================] - 0s 238ms/step 1/1 [==============================] - 0s 90ms/step 1/1 [==============================] - 0s 235ms/step 1/1 [==============================] - 0s 118ms/step 1/1 [==============================] - 0s 247ms/step 1/1 [==============================] - 0s 229ms/step 1/1 [==============================] - 0s 116ms/step 1/1 [==============================] - 0s 196ms/step 1/1 [==============================] - 0s 203ms/step 1/1 [==============================] - 0s 226ms/step 1/1 [==============================] - 0s 213ms/step 1/1 [==============================] - 0s 105ms/step 1/1 [==============================] - 0s 225ms/step 1/1 [==============================] - 0s 227ms/step 1/1 [==============================] - 0s 207ms/step 1/1 [==============================] - 0s 106ms/step 1/1 [==============================] - 0s 221ms/step 1/1 [==============================] - 0s 90ms/step 1/1 [==============================] - 0s 195ms/step 1/1 [==============================] - 0s 110ms/step 1/1 [==============================] - 0s 207ms/step 1/1 [==============================] - 0s 113ms/step 1/1 [==============================] - 0s 202ms/step 1/1 [==============================] - 0s 209ms/step 1/1 [==============================] - 0s 208ms/step 1/1 [==============================] - 0s 104ms/step 1/1 [==============================] - 0s 223ms/step 1/1 [==============================] - 0s 218ms/step 1/1 [==============================] - 0s 208ms/step 1/1 [==============================] - 0s 216ms/step 1/1 [==============================] - 0s 219ms/step 1/1 [==============================] - 0s 210ms/step 1/1 [==============================] - 0s 211ms/step 1/1 [==============================] - 0s 209ms/step 1/1 [==============================] - 0s 112ms/step 1/1 [==============================] - 0s 198ms/step 1/1 [==============================] - 0s 218ms/step 1/1 [==============================] - 0s 109ms/step 1/1 [==============================] - 0s 211ms/step 1/1 [==============================] - 0s 209ms/step 1/1 [==============================] - 0s 218ms/step 1/1 [==============================] - 0s 106ms/step 1/1 [==============================] - 0s 214ms/step 1/1 [==============================] - 0s 107ms/step 1/1 [==============================] - 0s 213ms/step 1/1 [==============================] - 0s 102ms/step 1/1 [==============================] - 0s 205ms/step 1/1 [==============================] - 0s 115ms/step 1/1 [==============================] - 0s 225ms/step 1/1 [==============================] - 0s 113ms/step 1/1 [==============================] - 0s 199ms/step 1/1 [==============================] - 0s 204ms/step 1/1 [==============================] - 0s 226ms/step 1/1 [==============================] - 0s 223ms/step 1/1 [==============================] - 0s 218ms/step 1/1 [==============================] - 0s 211ms/step 1/1 [==============================] - 0s 103ms/step 1/1 [==============================] - 0s 217ms/step 1/1 [==============================] - 0s 111ms/step 1/1 [==============================] - 0s 192ms/step 1/1 [==============================] - 0s 211ms/step 1/1 [==============================] - 0s 212ms/step 1/1 [==============================] - 0s 107ms/step 1/1 [==============================] - 0s 197ms/step 1/1 [==============================] - 0s 206ms/step 1/1 [==============================] - 0s 220ms/step 1/1 [==============================] - 0s 118ms/step 1/1 [==============================] - 0s 224ms/step 1/1 [==============================] - 0s 221ms/step 1/1 [==============================] - 0s 106ms/step 1/1 [==============================] - 0s 204ms/step 1/1 [==============================] - 0s 112ms/step 1/1 [==============================] - 0s 198ms/step 1/1 [==============================] - 0s 198ms/step 1/1 [==============================] - 0s 204ms/step 1/1 [==============================] - 0s 198ms/step 1/1 [==============================] - 0s 175ms/step 1/1 [==============================] - 0s 105ms/step 1/1 [==============================] - 0s 213ms/step 1/1 [==============================] - 0s 189ms/step 1/1 [==============================] - 0s 223ms/step 1/1 [==============================] - 0s 111ms/step 1/1 [==============================] - 0s 227ms/step 1/1 [==============================] - 0s 180ms/step 1/1 [==============================] - 0s 229ms/step 1/1 [==============================] - 0s 99ms/step 1/1 [==============================] - 0s 195ms/step 1/1 [==============================] - 0s 195ms/step 1/1 [==============================] - 0s 233ms/step 1/1 [==============================] - 0s 198ms/step 1/1 [==============================] - 0s 166ms/step 1/1 [==============================] - 0s 82ms/step 1/1 [==============================] - 0s 178ms/step 1/1 [==============================] - 0s 102ms/step 1/1 [==============================] - 0s 182ms/step 1/1 [==============================] - 0s 191ms/step 1/1 [==============================] - 0s 100ms/step 1/1 [==============================] - 0s 185ms/step 1/1 [==============================] - 0s 101ms/step 1/1 [==============================] - 0s 221ms/step 1/1 [==============================] - 0s 206ms/step 1/1 [==============================] - 0s 203ms/step 1/1 [==============================] - 0s 213ms/step 1/1 [==============================] - 0s 105ms/step 1/1 [==============================] - 0s 202ms/step 1/1 [==============================] - 0s 229ms/step 1/1 [==============================] - 0s 106ms/step 1/1 [==============================] - 0s 216ms/step 1/1 [==============================] - 0s 222ms/step 1/1 [==============================] - 0s 215ms/step 1/1 [==============================] - 0s 222ms/step 1/1 [==============================] - 0s 109ms/step 1/1 [==============================] - 0s 205ms/step 1/1 [==============================] - 0s 118ms/step 1/1 [==============================] - 0s 191ms/step 1/1 [==============================] - 0s 229ms/step 1/1 [==============================] - 0s 202ms/step 1/1 [==============================] - 0s 223ms/step 1/1 [==============================] - 0s 108ms/step 1/1 [==============================] - 0s 201ms/step 1/1 [==============================] - 0s 202ms/step 1/1 [==============================] - 0s 119ms/step 1/1 [==============================] - 0s 189ms/step 1/1 [==============================] - 0s 231ms/step 1/1 [==============================] - 0s 103ms/step 1/1 [==============================] - 0s 211ms/step 1/1 [==============================] - 0s 192ms/step 1/1 [==============================] - 0s 201ms/step 1/1 [==============================] - 0s 103ms/step 1/1 [==============================] - 0s 216ms/step 1/1 [==============================] - 0s 105ms/step 1/1 [==============================] - 0s 191ms/step 1/1 [==============================] - 0s 195ms/step 1/1 [==============================] - 0s 107ms/step 1/1 [==============================] - 0s 192ms/step 1/1 [==============================] - 0s 98ms/step 1/1 [==============================] - 0s 215ms/step 1/1 [==============================] - 0s 188ms/step 1/1 [==============================] - 0s 121ms/step 1/1 [==============================] - 0s 225ms/step 1/1 [==============================] - 0s 118ms/step 1/1 [==============================] - 0s 205ms/step 1/1 [==============================] - 0s 200ms/step 1/1 [==============================] - 0s 110ms/step 1/1 [==============================] - 0s 227ms/step 1/1 [==============================] - 0s 190ms/step 1/1 [==============================] - 0s 148ms/step 1/1 [==============================] - 0s 110ms/step 1/1 [==============================] - 0s 207ms/step 1/1 [==============================] - 0s 214ms/step 1/1 [==============================] - 0s 231ms/step 1/1 [==============================] - 0s 200ms/step 1/1 [==============================] - 0s 113ms/step 1/1 [==============================] - 0s 222ms/step 1/1 [==============================] - 0s 114ms/step 1/1 [==============================] - 0s 212ms/step 1/1 [==============================] - 0s 213ms/step 1/1 [==============================] - 0s 109ms/step 1/1 [==============================] - 0s 210ms/step 1/1 [==============================] - 0s 114ms/step 1/1 [==============================] - 0s 217ms/step 1/1 [==============================] - 0s 217ms/step 1/1 [==============================] - 0s 115ms/step 1/1 [==============================] - 0s 221ms/step 1/1 [==============================] - 0s 208ms/step 1/1 [==============================] - 0s 232ms/step 1/1 [==============================] - 0s 112ms/step 1/1 [==============================] - 0s 223ms/step 1/1 [==============================] - 0s 216ms/step 1/1 [==============================] - 0s 105ms/step 1/1 [==============================] - 0s 213ms/step 1/1 [==============================] - 0s 228ms/step 1/1 [==============================] - 0s 114ms/step 1/1 [==============================] - 0s 221ms/step 1/1 [==============================] - 0s 231ms/step 1/1 [==============================] - 0s 115ms/step 1/1 [==============================] - 0s 215ms/step 1/1 [==============================] - 0s 95ms/step 1/1 [==============================] - 0s 213ms/step 1/1 [==============================] - 0s 207ms/step 1/1 [==============================] - 0s 206ms/step 1/1 [==============================] - 0s 214ms/step 1/1 [==============================] - 0s 114ms/step 1/1 [==============================] - 0s 209ms/step 1/1 [==============================] - 0s 179ms/step 1/1 [==============================] - 0s 87ms/step 1/1 [==============================] - 0s 174ms/step 1/1 [==============================] - 0s 188ms/step 1/1 [==============================] - 0s 182ms/step 1/1 [==============================] - 0s 188ms/step 1/1 [==============================] - 0s 105ms/step 1/1 [==============================] - 0s 217ms/step 1/1 [==============================] - 0s 228ms/step 1/1 [==============================] - 0s 114ms/step 1/1 [==============================] - 0s 197ms/step 1/1 [==============================] - 0s 211ms/step 1/1 [==============================] - 0s 199ms/step 1/1 [==============================] - 0s 107ms/step 1/1 [==============================] - 0s 205ms/step 1/1 [==============================] - 0s 235ms/step 1/1 [==============================] - 0s 107ms/step 1/1 [==============================] - 0s 211ms/step 1/1 [==============================] - 0s 194ms/step 1/1 [==============================] - 0s 210ms/step 1/1 [==============================] - 0s 106ms/step 1/1 [==============================] - 0s 213ms/step 1/1 [==============================] - 0s 226ms/step 1/1 [==============================] - 0s 222ms/step 1/1 [==============================] - 0s 108ms/step 1/1 [==============================] - 0s 218ms/step 1/1 [==============================] - 0s 223ms/step 1/1 [==============================] - 0s 226ms/step 1/1 [==============================] - 0s 209ms/step 1/1 [==============================] - 0s 112ms/step 1/1 [==============================] - 0s 206ms/step 1/1 [==============================] - 0s 226ms/step 1/1 [==============================] - 0s 248ms/step 1/1 [==============================] - 0s 118ms/step 1/1 [==============================] - 0s 209ms/step 1/1 [==============================] - 0s 217ms/step 1/1 [==============================] - 0s 215ms/step 1/1 [==============================] - 0s 208ms/step 1/1 [==============================] - 0s 106ms/step 1/1 [==============================] - 0s 222ms/step 1/1 [==============================] - 0s 109ms/step 1/1 [==============================] - 0s 207ms/step 1/1 [==============================] - 0s 135ms/step 1/1 [==============================] - 0s 212ms/step 1/1 [==============================] - 0s 215ms/step 1/1 [==============================] - 0s 213ms/step 1/1 [==============================] - 0s 221ms/step 1/1 [==============================] - 0s 209ms/step 1/1 [==============================] - 0s 92ms/step 1/1 [==============================] - 0s 231ms/step 1/1 [==============================] - 0s 211ms/step 1/1 [==============================] - 0s 225ms/step 1/1 [==============================] - 0s 221ms/step 1/1 [==============================] - 0s 107ms/step 1/1 [==============================] - 0s 210ms/step 1/1 [==============================] - 0s 203ms/step 1/1 [==============================] - 0s 235ms/step 1/1 [==============================] - 0s 224ms/step 1/1 [==============================] - 0s 90ms/step 1/1 [==============================] - 0s 226ms/step 1/1 [==============================] - 0s 207ms/step 1/1 [==============================] - 0s 108ms/step 1/1 [==============================] - 0s 224ms/step 1/1 [==============================] - 0s 208ms/step 1/1 [==============================] - 0s 101ms/step 1/1 [==============================] - 0s 199ms/step 1/1 [==============================] - 0s 112ms/step 1/1 [==============================] - 0s 221ms/step 1/1 [==============================] - 0s 214ms/step 1/1 [==============================] - 0s 243ms/step 1/1 [==============================] - 0s 228ms/step 1/1 [==============================] - 0s 113ms/step 1/1 [==============================] - 0s 199ms/step 1/1 [==============================] - 0s 218ms/step 1/1 [==============================] - 0s 215ms/step 1/1 [==============================] - 0s 199ms/step 1/1 [==============================] - 0s 115ms/step 1/1 [==============================] - 0s 221ms/step 1/1 [==============================] - 0s 214ms/step 1/1 [==============================] - 0s 119ms/step 1/1 [==============================] - 0s 208ms/step 1/1 [==============================] - 0s 113ms/step 1/1 [==============================] - 0s 218ms/step 1/1 [==============================] - 0s 106ms/step 1/1 [==============================] - 0s 221ms/step 1/1 [==============================] - 0s 229ms/step 1/1 [==============================] - 0s 209ms/step 1/1 [==============================] - 0s 203ms/step 1/1 [==============================] - 0s 112ms/step 1/1 [==============================] - 0s 209ms/step 1/1 [==============================] - 0s 109ms/step 1/1 [==============================] - 0s 238ms/step 1/1 [==============================] - 0s 113ms/step 1/1 [==============================] - 0s 205ms/step 1/1 [==============================] - 0s 106ms/step 1/1 [==============================] - 0s 210ms/step 1/1 [==============================] - 0s 203ms/step 1/1 [==============================] - 0s 108ms/step 1/1 [==============================] - 0s 217ms/step 1/1 [==============================] - 0s 101ms/step 1/1 [==============================] - 0s 200ms/step 1/1 [==============================] - 0s 221ms/step 1/1 [==============================] - 0s 108ms/step 1/1 [==============================] - 0s 214ms/step 1/1 [==============================] - 0s 225ms/step 1/1 [==============================] - 0s 226ms/step 1/1 [==============================] - 0s 201ms/step 1/1 [==============================] - 0s 209ms/step 1/1 [==============================] - 0s 205ms/step 1/1 [==============================] - 0s 113ms/step 1/1 [==============================] - 0s 173ms/step 1/1 [==============================] - 0s 173ms/step 1/1 [==============================] - 0s 200ms/step 1/1 [==============================] - 0s 219ms/step 1/1 [==============================] - 0s 209ms/step 1/1 [==============================] - 0s 201ms/step 1/1 [==============================] - 0s 115ms/step 1/1 [==============================] - 0s 193ms/step 1/1 [==============================] - 0s 208ms/step 1/1 [==============================] - 0s 216ms/step 1/1 [==============================] - 0s 204ms/step 1/1 [==============================] - 0s 206ms/step 1/1 [==============================] - 0s 220ms/step 1/1 [==============================] - 0s 189ms/step 1/1 [==============================] - 0s 195ms/step 1/1 [==============================] - 0s 109ms/step 1/1 [==============================] - 0s 218ms/step 1/1 [==============================] - 0s 198ms/step 1/1 [==============================] - 0s 196ms/step 1/1 [==============================] - 0s 207ms/step 1/1 [==============================] - 0s 196ms/step 1/1 [==============================] - 0s 206ms/step 1/1 [==============================] - 0s 215ms/step 1/1 [==============================] - 0s 111ms/step 1/1 [==============================] - 0s 223ms/step 1/1 [==============================] - 0s 103ms/step 1/1 [==============================] - 0s 219ms/step 1/1 [==============================] - 0s 97ms/step 1/1 [==============================] - 0s 207ms/step 1/1 [==============================] - 0s 213ms/step 1/1 [==============================] - 0s 207ms/step 1/1 [==============================] - 0s 224ms/step 1/1 [==============================] - 0s 227ms/step 1/1 [==============================] - 0s 112ms/step 1/1 [==============================] - 0s 209ms/step 1/1 [==============================] - 0s 104ms/step 1/1 [==============================] - 0s 201ms/step 1/1 [==============================] - 0s 107ms/step 1/1 [==============================] - 0s 200ms/step 1/1 [==============================] - 0s 207ms/step 1/1 [==============================] - 0s 195ms/step 1/1 [==============================] - 0s 201ms/step 1/1 [==============================] - 0s 103ms/step 1/1 [==============================] - 0s 206ms/step 1/1 [==============================] - 0s 107ms/step 1/1 [==============================] - 0s 207ms/step 1/1 [==============================] - 0s 204ms/step 1/1 [==============================] - 0s 117ms/step 1/1 [==============================] - 0s 208ms/step 1/1 [==============================] - 0s 209ms/step 1/1 [==============================] - 0s 217ms/step 1/1 [==============================] - 0s 206ms/step 1/1 [==============================] - 0s 229ms/step 1/1 [==============================] - 0s 230ms/step 1/1 [==============================] - 0s 104ms/step 1/1 [==============================] - 0s 206ms/step 1/1 [==============================] - 0s 228ms/step 1/1 [==============================] - 0s 225ms/step 1/1 [==============================] - 0s 107ms/step 1/1 [==============================] - 0s 238ms/step 1/1 [==============================] - 0s 212ms/step 1/1 [==============================] - 0s 212ms/step 1/1 [==============================] - 0s 215ms/step 1/1 [==============================] - 0s 102ms/step 1/1 [==============================] - 0s 206ms/step 1/1 [==============================] - 0s 212ms/step 1/1 [==============================] - 0s 114ms/step 1/1 [==============================] - 0s 199ms/step 1/1 [==============================] - 0s 103ms/step 1/1 [==============================] - 0s 208ms/step 1/1 [==============================] - 0s 209ms/step 1/1 [==============================] - 0s 104ms/step 1/1 [==============================] - 0s 193ms/step 1/1 [==============================] - 0s 224ms/step 1/1 [==============================] - 0s 215ms/step 1/1 [==============================] - 0s 109ms/step 1/1 [==============================] - 0s 214ms/step 1/1 [==============================] - 0s 112ms/step 1/1 [==============================] - 0s 229ms/step 1/1 [==============================] - 0s 99ms/step 1/1 [==============================] - 0s 213ms/step 1/1 [==============================] - 0s 231ms/step 1/1 [==============================] - 0s 109ms/step 1/1 [==============================] - 0s 208ms/step 1/1 [==============================] - 0s 101ms/step 1/1 [==============================] - 0s 206ms/step 1/1 [==============================] - 0s 110ms/step 1/1 [==============================] - 0s 238ms/step 1/1 [==============================] - 0s 227ms/step 1/1 [==============================] - 0s 238ms/step 1/1 [==============================] - 0s 116ms/step 1/1 [==============================] - 0s 245ms/step 1/1 [==============================] - 0s 232ms/step 1/1 [==============================] - 0s 231ms/step 1/1 [==============================] - 0s 227ms/step 1/1 [==============================] - 0s 216ms/step 1/1 [==============================] - 0s 222ms/step 1/1 [==============================] - 0s 224ms/step 1/1 [==============================] - 0s 230ms/step 1/1 [==============================] - 0s 208ms/step 1/1 [==============================] - 0s 104ms/step 1/1 [==============================] - 0s 208ms/step 1/1 [==============================] - 0s 196ms/step 1/1 [==============================] - 0s 212ms/step 1/1 [==============================] - 0s 211ms/step 1/1 [==============================] - 0s 199ms/step 1/1 [==============================] - 0s 209ms/step 1/1 [==============================] - 0s 106ms/step 1/1 [==============================] - 0s 213ms/step 1/1 [==============================] - 0s 95ms/step 1/1 [==============================] - 0s 203ms/step 1/1 [==============================] - 0s 201ms/step 1/1 [==============================] - 0s 207ms/step 1/1 [==============================] - 0s 115ms/step 1/1 [==============================] - 0s 214ms/step 1/1 [==============================] - 0s 117ms/step 1/1 [==============================] - 0s 208ms/step 1/1 [==============================] - 0s 215ms/step 1/1 [==============================] - 0s 215ms/step 1/1 [==============================] - 0s 205ms/step 1/1 [==============================] - 0s 207ms/step 1/1 [==============================] - 0s 201ms/step 1/1 [==============================] - 0s 201ms/step 1/1 [==============================] - 0s 214ms/step 1/1 [==============================] - 0s 229ms/step 1/1 [==============================] - 0s 116ms/step 1/1 [==============================] - 0s 204ms/step 1/1 [==============================] - 0s 233ms/step 1/1 [==============================] - 0s 229ms/step 1/1 [==============================] - 0s 105ms/step 1/1 [==============================] - 0s 216ms/step 1/1 [==============================] - 0s 207ms/step 1/1 [==============================] - 0s 208ms/step 1/1 [==============================] - 0s 227ms/step 1/1 [==============================] - 0s 113ms/step 1/1 [==============================] - 0s 208ms/step 1/1 [==============================] - 0s 82ms/step 1/1 [==============================] - 0s 220ms/step 1/1 [==============================] - 0s 99ms/step 1/1 [==============================] - 0s 184ms/step 1/1 [==============================] - 0s 200ms/step 1/1 [==============================] - 0s 113ms/step 1/1 [==============================] - 0s 220ms/step 1/1 [==============================] - 0s 190ms/step 1/1 [==============================] - 0s 112ms/step 1/1 [==============================] - 0s 213ms/step 1/1 [==============================] - 0s 208ms/step 1/1 [==============================] - 0s 106ms/step 1/1 [==============================] - 0s 209ms/step 1/1 [==============================] - 0s 216ms/step 1/1 [==============================] - 0s 110ms/step 1/1 [==============================] - 0s 210ms/step 1/1 [==============================] - 0s 211ms/step 1/1 [==============================] - 0s 98ms/step 1/1 [==============================] - 0s 209ms/step 1/1 [==============================] - 0s 102ms/step 1/1 [==============================] - 0s 214ms/step 1/1 [==============================] - 0s 98ms/step 1/1 [==============================] - 0s 207ms/step 1/1 [==============================] - 0s 218ms/step 1/1 [==============================] - 0s 208ms/step 1/1 [==============================] - 0s 115ms/step 1/1 [==============================] - 0s 208ms/step 1/1 [==============================] - 0s 205ms/step 1/1 [==============================] - 0s 190ms/step 1/1 [==============================] - 0s 198ms/step
# creating a dataframe for the result
df_pred = pd.DataFrame({'image_path': image_id,'predicted_mask': mask,'has_mask': has_mask})
df_pred
| image_path | predicted_mask | has_mask | |
|---|---|---|---|
| 0 | TCGA_DU_5872_19950223/TCGA_DU_5872_19950223_67... | No mask | 0 |
| 1 | TCGA_FG_A4MT_20020212/TCGA_FG_A4MT_20020212_3.tif | No mask | 0 |
| 2 | TCGA_DU_A5TW_19980228/TCGA_DU_A5TW_19980228_29... | [[[[6.579732e-06], [5.0200333e-07], [1.8209814... | 1 |
| 3 | TCGA_FG_7637_20000922/TCGA_FG_7637_20000922_17... | [[[[1.3748944e-06], [5.6489725e-08], [3.797600... | 1 |
| 4 | TCGA_CS_6666_20011109/TCGA_CS_6666_20011109_25... | [[[[4.1414055e-06], [1.4924562e-06], [7.196988... | 1 |
| ... | ... | ... | ... |
| 585 | TCGA_HT_A61A_20000127/TCGA_HT_A61A_20000127_23... | [[[[3.941212e-06], [1.2909313e-06], [7.090385e... | 1 |
| 586 | TCGA_DU_5872_19950223/TCGA_DU_5872_19950223_69... | No mask | 0 |
| 587 | TCGA_DU_6404_19850629/TCGA_DU_6404_19850629_24... | No mask | 0 |
| 588 | TCGA_DU_8165_19970205/TCGA_DU_8165_19970205_7.tif | No mask | 0 |
| 589 | TCGA_DU_6404_19850629/TCGA_DU_6404_19850629_53... | No mask | 0 |
590 rows × 3 columns
# Merge the dataframe containing predicted results with the original test data.
df_pred = test.merge(df_pred, on = 'image_path')
df_pred.head()
| image_path | mask_path | mask | predicted_mask | has_mask | |
|---|---|---|---|---|---|
| 0 | TCGA_DU_5872_19950223/TCGA_DU_5872_19950223_67... | TCGA_DU_5872_19950223/TCGA_DU_5872_19950223_67... | 0 | No mask | 0 |
| 1 | TCGA_FG_A4MT_20020212/TCGA_FG_A4MT_20020212_3.tif | TCGA_FG_A4MT_20020212/TCGA_FG_A4MT_20020212_3_... | 0 | No mask | 0 |
| 2 | TCGA_DU_A5TW_19980228/TCGA_DU_A5TW_19980228_29... | TCGA_DU_A5TW_19980228/TCGA_DU_A5TW_19980228_29... | 0 | [[[[6.579732e-06], [5.0200333e-07], [1.8209814... | 1 |
| 3 | TCGA_FG_7637_20000922/TCGA_FG_7637_20000922_17... | TCGA_FG_7637_20000922/TCGA_FG_7637_20000922_17... | 1 | [[[[1.3748944e-06], [5.6489725e-08], [3.797600... | 1 |
| 4 | TCGA_CS_6666_20011109/TCGA_CS_6666_20011109_25... | TCGA_CS_6666_20011109/TCGA_CS_6666_20011109_25... | 0 | [[[[4.1414055e-06], [1.4924562e-06], [7.196988... | 1 |
count = 0
fig, axs = plt.subplots(10, 5, figsize=(30, 50))
for i in range(len(df_pred)):
if df_pred['has_mask'][i] == 1 and count < 10:
# read the images and convert them to RGB format
img = io.imread(df_pred.image_path[i])
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
axs[count][0].title.set_text("Brain MRI")
axs[count][0].imshow(img)
# Obtain the mask for the image
mask = io.imread(df_pred.mask_path[i])
axs[count][1].title.set_text("Original Mask")
axs[count][1].imshow(mask)
# Obtain the predicted mask for the image
predicted_mask = np.asarray(df_pred.predicted_mask[i])[0].squeeze().round()
axs[count][2].title.set_text("AI Predicted Mask")
axs[count][2].imshow(predicted_mask)
# Apply the mask to the image 'mask==255'
img[mask == 255] = (255, 0, 0)
axs[count][3].title.set_text("MRI with Original Mask (Ground Truth)")
axs[count][3].imshow(img)
img_ = io.imread(df_pred.image_path[i])
img_ = cv2.cvtColor(img_, cv2.COLOR_BGR2RGB)
img_[predicted_mask == 1] = (0, 255, 0)
axs[count][4].title.set_text("MRI with AI Predicted Mask")
axs[count][4].imshow(img_)
count += 1
fig.tight_layout()
